Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics, Autonomous mobile robots and manipulators
Abstract: The walking control of bipedal robots poses challenges due to inherent coupling among the robot's degrees of freedom. This paper introduces an approach to address this challenge by using decoupled control in the sagittal and frontal planes. The proposed control method takes advantage of Hybrid Zero Dynamics and Hybrid-Linear Inverted Pendulum for sagittal and frontal plane dynamics, respectively. The hybrid controller is successfully validated on a bipedal robot RobBIE, whose torso inertia is relatively high and if not adequately controlled can easily violate the point mass assumption in many reduced-order model based walking controllers developed previously. With the help of full-model based Hybrid Zero Dynamics, the robot can achieve stable walking behaviors at different velocities and adapt to various terrains and even moderate disturbances.

Keywords: Soft robotics and liquid-metal robotics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Smart sensors and actuators
Abstract: Soft crawling robots have gained attention due to their ability to navigate through complex environments. In this work, we present a multi-degrees-of-freedom (multi-DOF) crawling robot driven by a square planar dielectric elastomer actuator. The robot is equipped with four independent distributed electrodes, enabling it to achieve straight driving and turning by controlling the activation of each electrode. The results show that when excite three electrodes, the robot achieves a peak velocity of 60mm/s during straight crawling, which is 1.03 times of body length. Additionally, by exciting two adjacent electrodes, the robot can turn at an angular velocity of 9.16°/s. The robot proposed in this work has a simple structure and is promising for applications such as detection in narrow space.

Keywords: Smart sensors and actuators, New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: In recent years, there has been significant advancement in the technological development of shape memory alloy (SMA) actuators, resulting in their increased maturity and wider applicability. Concurrently, the industry's demand for compact, rapid-response actuators with high precision has seen a notable rise. In light of this, this study presents a novel SMA actuator utilizing Ni-Ti shape memory alloy (Ni-Ti SMA) wire, driven by the concept of SMA wire heating-induced contraction. The proposed actuator serves a dual purpose: measuring the intrinsic self-sensing attributes of the SMA wire and accurately ascertaining displacement alterations through the phase transition of the SMA wire, thereby precisely determining positional changes. By capitalizing on the resistance-displacement mapping relationship during the SMA wire's phase transformation, autonomous self-sensing control is achieved without the need for external sensors. To meet the imperatives of rapid response, superior precision, and minimal form factor for the SMA wire, the actuator is ingeniously integrated to incorporate an auxiliary position sensor and a self-resetting loading mechanism. This integration significantly downsizes the overall dimensions of the SMA actuator, aligning it with the criteria of compactness while retaining performance.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics, Soft robotics and liquid-metal robotics
Abstract: Pipeline inspection poses challenges that require robots to navigate complex environments. However, existing pipeline robots with the rigid bodies often face limitations in terms of efficiency, flexibility, and adaptability considering the significant size variations and complex terrains in pipelines. This paper introduces a novel design of wheel pipeline robots that addresses these limitations by incorporating the Origami Anisotropic Stiffness Structure (OASS). Inspired by the desert iguana's skin, the OASS offers a unique combination of rigidity and flexibility in different directions, making it an ideal skeletal framework for the robot with only 146 g weight. By integrating the OASS with PLA and resin materials, we enhance the robot's support and enable it to effectively navigate diverse pipeline configurations. Experimental results demonstrate the robot's capability to maneuver through pipelines of different diameters, U-shaped turns, obstacles, and even vertical sections. This research provides a promising solution for inspection tasks in complex pipeline systems.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: Small-scale robots are widely used in real-world rescue missions, but their mobility and movement range are still limited. One solution to improve their adaptability to complex environments is to introduce jumping and gliding strategies into the robot design. In this paper, we developed a small-scale locust-inspired robot capable of active (propeller-driven) gliding after launching from the ground, which had a body length of 19.1 cm and a weight of 97 g. On the basis of the locust’s musculoskeletal model, the jumping system was designed by a six-bar mechanism which can amplify the power. To improve the lift in gliding phase while reducing the draft in jumping phase, the gliding system with a folding wing and a front propeller was proposed corresponding to the fixed-wing principle. The results obtained through a series of experimental tests reveal that the robot achieves a jumping height of 0.15 m and a passive gliding distance of 1.5 m, which has a glide ratio of 1.13. Remarkably, actuating by a propeller, the robot can jump up to a height of 0.20 m, covering a gliding distance of 2.9 m. It is worth noting that the glide ratio of the robot improves by 91.2% in the propeller-driven jump-gliding mode.

Keywords: New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics, Robotics in intelligent manufacturing
Abstract: The authors proposed family of novel six-dof parallel robots comprised of two sub-parallel mechanisms that can directly tilt the MP around x and y axis, and extend the z-axis rotational range by a simple and lightweight one-DOF additional rotation mechanism. This paper reports on developing prototypes and verifying the operations of parallel robots. The notable feature of this research is to construct a single re-configurable prototype machine and controller for verification of the operation of three types of parallel robots. Firstly, the kinematical structures of our proposed robots, which we dubbed PEN2 family, are briefly introduced. Next, joints of the limbs on the arm side and MP side are replaced with equivalent detachable spherical joints for the re-configurable prototype design. Then a novel eazy to re-configurable prototype mechanism is designed for the modified PEN2 family. In addition, the development of a Simulink-based controller for the PEN2 family that can be easily re-configured is described. Motions of the PEN2 family were verified using the proposed re-configurable prototypes and controllers.

Keywords: Human-robot interaction, Multi-robot systems, swarm robots, and collaborative robots, Smart sensors and actuators
Abstract: Utilizing the hand-based input with haptic feedback to interact with a robot team has attracted significant interest because of its high situational awareness. The main disadvantages of this interaction include insufficient teleoperation tasks due to the mechanical limitation of the haptic device, cognitive load caused by the contradiction between the highly-abstracted control command and the low-level control input, and added physical fatigue caused by intensive input operations. In this paper, we propose a novel interaction scheme for bilaterally teleoperating a group of mobile robots by combining the eye-gaze input with the hand input. In the scheme the controller of the robot team is divided into main tasks and subtasks. The main tasks are finished by the master devices. The haptic device is utilized to receive the virtual force feedback and control the formation size and velocity. The eye tracker is used to generate the formation transition commands. The subtasks including obstacle avoidance and formation keeping are accomplished autonomously by the robot team. The experimental results show that the scheme is more acceptable, intuitive, and efficient compared with the traditional bilateral teleoperation scheme without gaze input.

Keywords: Multi-robot systems, swarm robots, and collaborative robots
Abstract: Formation control is an essential research topic in multi-agent systems (MAS), while the convergence speed of formation is critically important for applications with real-time performance requirements, such as rescuing tasks. However, there is still a lack of effective methods for practically usable formation control with controllable convergence speed. This paper introduces a novel Laplacian function-based approach to enhance the convergence speed of MAS in formation control. By utilizing the Laplacian matrix of the communication graph, eigenvalues are mapped to desired positions, thereby improving the convergence speed of the formation process. Additionally, this approach enables estimation and manipulation of the convergence speed, offering flexibility and adaptability to meet application-specific requirements. The proposed method is experimentally validated through multiple quadrotors, demonstrating its effectiveness and practical feasibility. Experimental results show that the formation convergence speed can be well controlled by appropriately designing the Laplacian functions.

Keywords: Multi-sensor data fusion and sensor networks
Abstract: In this article, we propose a framework for tightly coupled lidar inertial odometry, which can achieve highly accuracy in real-time ego-motion estimation and map building of the robot. The estimated motion from inertial measurement unit (IMU) preintegration de-skews point clouds. Then, an initialization method based on MAP estimation is used to align the Lidar frame with the world frame and provide reliable initial values for the system. In order to make the scan matching more accurate, we track the state of each frame, and transform the feature points to the reference frame to build a local map for matching with key frames. To ensure high performance in real-time, an efficient sliding window was used to optimize the keyframes we chose. After obtaining the optimized pose of the keyframe, we further optimize the pose of the normal frame by pose constraints. In addition, loop detection is integrated into our algorithm framework to eliminate accumulated errors, making our algorithm to obtain a globally consistent estimate. The experiment results demonstrate that our method can achieve high accuracy in a variety of indoor and outdoor environments even under fast motion conditions or with insufficient features.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Robotics in intelligent manufacturing
Abstract: The performance of multi-robot systems heavily relies on efficient task allocation and motion coordination. However, for a group of a large number of robots, finding the optimal solution is inevitably time-consuming and may become impossible. Recognizing that tasks vary in their impact on system performance, our main idea is to identify their critical subset that significantly influences the entire system, and enhance task allocation efficiency by optimally planning critical tasks while distributing the remaining tasks randomly or with simple strategies. We call this approach Selective Multirobot Task Planning (SMTP), which contributes to significantly reducing the computational requirements and solution time, and in the meanwhile, maintaining the system performance. In addition, by implementing a filtering mechanism based on the conditional expectation to eliminate less essential tasks, SMTP shows high extendability, maximizes task allocation efficiency, and balances computational efficiency and solution quality. Massive simulation and real experiments demonstrate that our algorithm decreases the computation time and maintains the properties of the base system. Large-scale experiments show that our approach only takes 11% computation time to reach 80% optimization objective and 94.8% traffic balance performance.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Multi-robot systems, swarm robots, and collaborative robots, Autonomous mobile robots and manipulators
Abstract: A blimp, called FishBlimp, that uses fins for propulsion was built and source localisation capability was added to it. The blimp's envelope is a spherical mylar foil balloon approximately 53 cm in diameter. To this, servos and fins were added in four equal positions along the equator of the envelope. A gondola was attached to the bottom with the rest of the hardware. Support for the blimp was added to the open source autopilot called ArduPilot. Blimp dynamics was also added to its simulation system called SITL. Then multiple-agent source localisation was performed within SITL, using a modified Particle Swarm Optimisation algorithm.

Keywords: Smart sensors and actuators, Soft robotics and liquid-metal robotics
Abstract: In recent years, flexible sensors have rapidly become an emerging research topic. Nevertheless, most existing flexible sensors focus primarily on strain and pressure or force. Curvature, as a crucial measurement parameter, especially for evaluating the posture of soft & rigid robots or humans, however, has been rarely studied relatively. Here, we propose an ultra-flexible liquid metal curvature sensor based on a double-layer multi-microchannel structural design. The sensor can measure the bending by electric resistance variations, instead of relying on elongations like most flexible curvature sensors used now. When the sensor is bent, the embedded microchannel filled with liquid metal becomes narrow under compressive stress, and then due to the linear increase in electric resistance of the liquid metal, the curvature can be effectively measured. Subsequently, we introduce an easy and low-cost manufacturing method for curvature sensor. Then a theoretical model is established to further verify the rationality and functionality of the sensor. Finally, we present experimental results to estimate the sensor characterization. It proves that the ultra-flexible liquid metal curvature sensor has a higher sensitivity within a wide range degree, very low hysteresis and is still robust in withstanding >800 repeated loading and unloading cycles.

Keywords: Applications of robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Soft robotics and liquid-metal robotics
Abstract: A soft inchworm robot with active friction control of wheels using soft Double-Network (DN) gel was developed and studied. The rotation of the robot’s wheel was controlled using brakes fabricated from DN gel. DN gel is a unique material that exhibits a frictional difference in surface upon voltage application. The directional crawling of the inchworm robot was possible through appropriate activation of DN gel brakes exploiting the property of surface friction. The brakes are in constant contact with the wheel’s axle and impress high and low friction upon voltage application to control the rotation. The control of rolling friction of the wheels results in improved locomotion capabilities and higher speeds compared to inchworm robots that employ sliding friction and body peristalsis. A mathematical model was developed to evaluate the theoretical displacement of the inchworm robot. The system is characterized by low power consumption, tough but soft brake material with good elasticity, low cost, and easy to fabricate with inherent compliance. The DN gel can be fabricated from special 3D printers which makes the whole robot completely 3D printable.

Keywords: Soft robotics and liquid-metal robotics, New theory and technology in robotics and biomimetics, Autonomous mobile robots and manipulators
Abstract: This paper presents analytical formulations for the kinematics and dynamics of continuum soft robotic arms using the Timoshenko beam theory (TBT). It has been shown that TBT is more comprehensive and accurate in compare to other beam models, and offers higher fidelity for modeling the flexible structures needed for control algorithm development and dynamic simulation. A soft robotic arm segment was considered with assumed planar motion where the dynamics of the soft segment was derived based on the TBT for two cases of solid and hollow cross-section. The material was modeled as elastic and the constitution laws was derived for the soft materials. The distributed pressure actuation was incorporated into the overall TBT-based dynamics of the continuum soft robotics arm with hollow structure. Numerical simulations were carried out for different cases with force and moment acting at the distal boundary condition and applied internal pressure.

Keywords: Soft robotics and liquid-metal robotics, Autonomous mobile robots and manipulators, New theory and technology in robotics and biomimetics
Abstract: This paper presents the dynamic modeling and trajectory tracking control of a shape morphing soft cylindrical robot capable of rolling and steering. Kinematics and dynamics of the robot were derived based on simplified assumptions and simulated via MATLAB to analyze the motion of the robot. A physical manifestation including a preliminary design and prototyping were demonstrated to develop the robot. A control architecture were conceived that decouples the robot’s dynamics into two cascaded subsystems, 1) a SISO rolling speed control and 2) a MIMO steering control. The feedback linearization control law and state feedback controller were derived for each subsystem where the asymptotic stability of the subsystems and the overall dynamical system were established. The controller performance in following the desired trajectories were examined in a series of simulation studies. The results showed successful demonstrations of the proposed shape-morphing steering mechanism for the robot’s locomotion.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Extensible Soft Robots (ESRs) have increasingly attracted attention, especially in tight and complex environ- ments, owing to their dexterity, and wide reachable workspaces relative to their volume. The existing actuation methods of ESR still suffer from challenges including large dimensions, expensive, complex modeling and control, and limited payload capacity. In this paper, we introduce a novel design for a lightweight, extensible and variable stiffness soft robot-based cable-driven actuation without backbone. The design composes of one section with three segments based on semi-octagon honeycomb structure, wherein the middle segment acts like a soft spring to provide extension feature. Additionally, the top and bottom segments have identical structure with two honeycomb patterns embedded within each other to improve the stiffness of the design. A design analysis is conducted to optimize the proposed structure with respect to stress and displacement. Additionally, a shape estimation approach is utilized to get accurate inference of the prototype’s shape based on data acquired from a low-cost Inertial Measurement Unit (IMU) and Motion Capture System (MCS), and constant curvature assumption. A series of experiments including workspace reachability, repeatability, shape approximation and payload analysis are carried out to validate the proposed design. The results show that the prototype exhibits good compression/extension ratio up to 42% and 50% relative to its normal length, respectively. Moreover, its payload capability reaches to 565 grams.

Keywords: Space robots, aerial robots, and underwater robots
Abstract: This paper proposes a method to increase the inspection speed of in-pipe robots in long, narrow, and complex pipes with long actuator response time including dead time and rise time due to long fluid tubes. The contributions of this paper include describing the novel approach to increase the moving speed of the peristaltic motion type in-pipe inspection robot with an endoskeleton enhancing propulsion and traction. Through the incorporation of a control method considering dead time, the developed peristaltic motion type robot improved its moving speed by 10.0% under standard pressure conditions. In addition, both theoretical and measured data displayed a similar speed enhancement trend. When integrating the high-pressure application method with the control method considering dead time, the resultant moving speed was found to be lower compared to the method without dead time consideration. This discrepancy can be attributed to the slower contraction response of pneumatic actuators relative to pressure response, leading to insufficient contraction during dead time superimposition. Analyzing the contraction response of the pneumatic tube and actuator used in this study revealed that the unit transitioned to the subsequent movement before achieving optimal contraction. However, the control method considering dead time under standard pressure conditions led to a promising 10.0% improvement in moving speed. This result indicates that the robot holds the potential for delivering both great force and speed in its operations. Experimental observations underscored the preference for applying dead time consideration to standard pressure application rather than high-pressure application methods when aiming to enhance performance.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: In this study, we conducted fundamental conveying experiments involving water and solid-liquid mixed fluids utilizing a peristaltic flexible pump. This pump was developed to imitate the peristaltic motion performed by the intestinal tracts of living organisms. This pump can convey a flexible path like an intestine because the piping itself expands and contracts flexibly in the long axial and radial directions. The pump consists of several identical units arranged in series. Each unit can be transported at a maximum bending angle of 25° per unit by employing a flexible bellows in its chassis. In this study, we experimented for transporting water and solid-liquid mixed fluids at angles of 0°, 45°, and 90° (Define 0° as the state where the pump is perpendicular to the ground) using the pump consists of five units. Experiments using this apparatus have shown that it is capable of transporting water and solid-liquid mixed fluids sufficiently at all of the above bending angles. The proposed system is expected to be applied in various fields, such as food, medicine, and extreme environments, because it can be transported regardless of the effects of gravity and transport paths.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Space robots, aerial robots, and underwater robots, Artificial intelligence in robotics
Abstract: This paper studies the modeling problems of a whole ornithopter in the Webots simulator. We introduce the quasi-steady aerodynamics and flapping mechanisms into the simulator. Then a trajectory-tracking project is carried out. With ROS2 integration, we have succeeded in designing a practical autopilot for flapping wing aerial vehicle research. The experimental results show that the ornithopter can not only reach steady flight with moderate thrust and attitude but also track the positional trajectory as expected.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: In order to provide more possibilities for replacing manual work in perilous high-altitude building operations, A novel mobile wall-climbing robot has been developed, named ClimBot-I. This robot is equipped with a lightweight, foldable, long-span mechanical arm that enables a wide range of functions. A highly reliable adsorption device with the design of a cyclone suction cup based on the Bernoulli principle is proposed. The dynamic modeling of the whole wall-climbing robot accompanying manipulator is completed. Through no-load and load experiments of the robot, the excellent adaptability for various wall surfaces and the ability to achieve stable adhesion are shown. The execute prescribed maneuvers on smooth substrates such as plastered walls are demonstrated. The robot can carry more than 3.6kg with the long-span operation manipulator on board and can move freely on the wall, which illustrate its broad application prospects.

Keywords: New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: The timing belt transmission offers numerous advantages for legged robots, including high efficiency, impact absorption and large range of joint motion. However, the transmission error under high load remains challenging to locomotion control and further applications of belt transmission. Traditional linear models cannot effectively model the belt deformation under a wide range of tension variations due to the nonlinearity. In this paper, we propose a model of the compensation for the belt transmission error based on the pretension and torque of the pully. The adopted approach bypasses the complexity of elaborate physical model derivations, yielding a non-linear model for transmission system errors through straightforward fitting. Based on the proposed model, an error compensation control is investigated and tested with an one-DoF leg prototype of legged robot. The alignment between experimental results and theoretical analysis demonstrates the accuracy of the modeling and the effectiveness of the error compensation control method. The proposed model provides a convenient and straightforward solution to effectively compensate for the belt transmission errors in legged robots.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics
Abstract: Legged robots have high mobility for application with uneven terrain. Bipedal robots require a lower number of actuators with less complication of control system. The five-bar mechanisms are widely applied in robotic legs to enable the leg movement to be actuated by the motors located on the robot body (instead of at the leg joints) to minimize the inertia of the moving legs. For attaining a desired foot position with respect to the two actuated joints of each leg, there are two non-trivial solutions obtained from the inverse kinematics. For both solutions, the parameters of the five-bar linkage are optimized using the derivative-free method by considering the energy consumption of the stance leg. Based on the optimized leg parameters, the zero-moment point (ZMP) along the foot support is derived by using the table-cart model. The ZMP generator for shifting the center of mass (COM) forward according to the stride is simulated with the dynamic model of the leg for predicting the motor torques varying in the stance phase and comparing between both solutions. Based on the optimal five-bar parameters, the small-scaled leg prototype was built, and the experiment was conducted to evaluate the trajectory and the actuation torque of each joint varying against the stride.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: The running of animals and humans often exhibits a spring-like leg behavior, which is abstractly explained by the Spring-Loaded Inverted Pendulum (SLIP) model. However, such an equivalent model neglects the nonlinear characteristics generated by the spring-like behavior localized at the joint level, leading to a substantial difference from those in the real system when analyzing locomotion stability and energy efficiency. The segmented leg model introduces the stiffness and rest angle of the virtual torsion spring at the joint into the dynamics to demonstrate the nonlinear relationship between the leg force and the leg compression. Due to the introduction of multiple parameters, it is of great significance to determine the optimal parameter combination. In this paper, we present a method to analyze the effects of the model parameters on the self-stability and energy efficiency. The nonlinear relationship between leg force and leg compression, and the hybrid dynamics of a two-segmented leg model are built, the apex return map is introduced to set up the self-stable constraints based on passive dynamics, and the effects of model parameters on the running energy efficiency are investigated via numerical simulation. The simulation results reveal that the highest energy efficiency is achieved when the stiffness is set to be the maximum value allowed and achievable, and the target running pattern is located at the fixed point corresponding to the largest angle to attack. The methodology to determine the model parameters is concluded based on the simulation results.

Keywords: Robotics in intelligent manufacturing, Robotic vision and image processing, Smart sensors and actuators
Abstract: As the global population grows exponentially, food consumption rises at a drastic rate. However, the limited availability of agricultural lands and inadequate traditional farming techniques have led to an emergence of alternatives such as hydroponic and aeroponic farming. Nonetheless, these systems still have limits, especially in terms of their constrained expansion when reliant on machinery, and the restricted vertical reach for human labor-driven harvesting. Therefore, we proposed an autonomous modular harvesting system for vertical farming of green oak lettuce. The structure is designed as an expandable cartesian gantry. With the initial base area of 40 x 40 cm, the system could be expanded to 120 cm in length and 240 cm in height, potentially hosting up to 96 lettuces. The system is also integrated with advanced computer vision to detect the presence of a plant and evaluate its readiness, primarily through color detection. Additionally, a specialized mechanism is installed to harvest the ripe lettuce. The system shows a notably high level of accuracy and impressive success rate in both detecting and harvesting.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics, Soft robotics and liquid-metal robotics
Abstract: In order to control the motion of a robot, a successful approach is to approximate the robot dynamics as a simplified model. However, the discrepancies between the actual mechanical properties of the robot and the simplified model will result in motion failure for the robot. To address this issue, this paper proposes a pneumatic-driven bipedal musculoskeletal robot that match the mechanistic properties of a simplified spring-loaded inverted pendulum (SLIP) model. The SLIP model is widely applied to robots because it exhibits passive stability and dynamic properties that are similar to human gaits. We designed a musculoskeletal biped robot with its center of mass concentrated in the small body near the hip joint, with low leg inertia based on the properties of the SLIP model. In addition, it it has been verified that the robot exhibits similar characteristics to the SLIP model through a sequential jumping experiment.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Space robots, aerial robots, and underwater robots
Abstract: The rock-climbing fish "glides" on the underwater wall at a maximum speed of 16 BL/s. Meanwhile, they are also sucker creatures that firmly attach themselves to the underwater wall under the impact of water currents. Previous studies have shown that the integration of suction-climbing functionality in these fish is mainly derived from the microstructure around the suction cups. In this paper, we found that the coordinated deformation of the full body of the rock-climbing fish has a positive effect on the motion of the suction-climbing mode by performing finite element method simulations of the airfoil of the rock-climbing fish in a fluid environment. We then present a novel single-motor-based design of a wire-driven robotic fish for bionic rock-climbing fish swimming, which is capable of realizing different propulsive states of the rock-climbing fish. Our findings provide support for the development of underwater suction-climbing robots based on the bionic rock-climbing fish design, encouraging further research in this area.

Keywords: New theory and technology in robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics
Abstract: In consideration of the poor locomotion ability of most traditional tensegrity robot, a novel tensegrity hopping robot powered by push-pull electromagnets was proposed with better locomotivity. It is able to conduct stable consecutive progressing action and turning action through hopping. This paper covers the structural design, theoretical modeling of the robot’s hopping process, as well as its self-righting analysis, simulation and experimental verification. The average moving speed and relative moving speed were measured to be 0.334m/s and 0.641 body length/s respectively, which obviously surpass the general moving ability of most traditional tensegrity robot. In addition, the robot was also validated to have good self-righting characteristic for stable consecutive hopping. Thus this robot can be potentially applied for search, rescue, detection, etc. in outdoor and unstructured environment.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Autonomous mobile robots and manipulators
Abstract: Abstract— In this paper, a bionic-based design of a series-parallel hybrid wheel-legged quadruped robot that can realize wheel-foot transformations is presented to achieve footed stable walking, and a method of measuring the static stability of the robot by the evaluation of the torque of the supporting leg rotor axis is proposed with the consideration of the robot's leg mass. Firstly, the whole robot position is solved; Secondly, the force on the robot during static walking is analyzed, and the stability of its static walking is also analyzed from the moment perspective; Finally, the stabilizing support region for static walking is solved, and the stabilizing support region for the robot in such a motion state is calculated. This paper provides high reference significance and value for the mechanism design and walking stability research of series-parallel hybrid wheel-legged quadruped robots.

Keywords: New theory and technology in robotics and biomimetics, Robotics in intelligent manufacturing
Abstract: Large tilt angle is required for parallel manipulators in many applications, this is a challenging issue in the field. In this paper, the optimum design of a 3-RCU parallel manipulator with 1T2R DoFs is carried out to realize the performance of large tilt angle output. The parameter-finiteness normalization method is used to build the parameter design space, and the motion/force transmission and constraint performance indices are used as the evaluation criterion. On these bases, the performance charts have been generated. Taking the constraint condition of achieving 45° tilt angle in all directions into consideration, an optimum region in the parameter design space has been derived and a group of optimized parameters is obtained. According to the results of optimum design, an CAD model of the manipulator is built. Based on perturbation method and principle of virtual work, a stiffness analytical model is established. Finally, the stiffness has been investigated, and the accuracy of the stiffness analytical model has been verified by comparing with the stiffness calculation using finite element analysis method. The work in this paper lays the foundation for the development of the manipulator.

Keywords: New theory and technology in robotics and biomimetics
Abstract: This paper proposed a line-symmetric double-centered 6R metamorphic mechanism with bifurcation properties. The mechanism is constructed based on elliptical geometric properties, which is a comprehensive design method for constructing novel metamorphic mechanisms through intuitive geometric figures. Firstly, the geometric constraints of the 6R metamorphic mechanism is revealed, and the parameter constraints and closed-loop kinematic equations of the 6R metamorphic mechanisms are deduced. Through the analysis of mechanism geometric constraints, four motion branches of the 6R metamorphic mechanism are obtained: collinear deployable motion branch MB1, spherical 4R motion branch MB2, spherical 4R motion branch MB3, and line-symmetrical Bricard motion branch MB4. Meanwhile, according to the screw theory, the analytical expressions and mobility of the constraint-screw systems under different motion branches are deduced. This led to the identification of the constraint and motion branch variations of the 6R metamorphic mechanism. Finally, the four serial link 4R singular configurations of the 6R metamorphic mechanism are specially studied, and the characteristics of the mutual reconfiguration of these four singular configurations are revealed.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Autonomous mobile robots and manipulators, Human-robot interaction
Abstract: This work proposes an emergency care system based on two robotic arms and task-specific end-effectors to be used in emergency care scenarios for life-saving pre-hospital treatment. The emergency scenario consists of a patient without access to human first aid and possibly in a remote, isolated location. Technological medical first aid is to be provided remotely through a proposed robotic emergency care system. The system will be transported to the patient via unmanned aerial or ground vehicle (UAV/UGV). In combination with independent interchangeable robotic end-effectors, telerobotic manipulation is the basis for all interventions and the focus of this work. End-effector tasks include clothing removal, oxygen administration, stopping critical bleeding, basic ultrasound diagnostics, and administration of emergency medication. This work shows the first realizations of end-effectors for oxygenation, tourniquet application, and administration of emergency medication. An end-effector for cutting clothing away via telerobotic manipulation in emergency situations was developed and tested on a synthetic mannequin with textiles from everyday clothing.

Keywords: Artificial intelligence in robotics, Robotics in intelligent manufacturing
Abstract: With the increasing emphasis on both quality and efficiency of automatic production, the motion stability and accuracy of a robot manipulator under heavy loads are becoming more and more important. It becomes imperative to optimize the dynamic performance of a robot manipulator for heavy-loads tasks, that is to improve the design of the mechanism and to optimize the mechanical properties of the whole system. In this study, we analyzed the dynamic model of a serial manipulator used for replacing disc-cutters in Tunnel Boring Machines and established an iterative algorithm to calculate the actuation torque for each joint during dynamic tasks. The dynamic manipulability index and the torque optimization index were proposed to evaluate the actuation requirement for the whole system. Then, a multi-objective optimization framework was established considering both kinematics and dynamics performance of the manipulator. The optimization was implemented using NSGA-II algorithm and we evaluate the optimize result from length compactness index and length balance index. The torque of the solution after optimization is obviously reduced and meet the performance index. The proposed optimization method can help improve operational precision and efficiency in constrained space.

Keywords: Robotics in intelligent manufacturing
Abstract: 摘要—随着增材制造技术的不断发展，越来越多的领域与3D打印集成，为其发展提供了新的机遇。本文提出了一种新型的移动式3D打印机器人，该机器人配合平面五杆机构克服了传统3D打印机的不动和工作范围受限等局限性。移动3D打印机器人具有独特的工作模式，利用平面五杆机构来跟踪平面模型的轨迹。机器人控制线性模型，实现顶出器的提升，从而完成制造过程。此外，当与移动平台配对时，机器人能够在x轴和y轴上无限延伸。在机器人上进行了轨迹模拟和工作演示，展示了其实现复杂轨迹和制造大型模型的!

Keywords: Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics
Abstract: To exactly detect the water ice of the South Pole of the moon, a lunar regolith in-situ analysis payload deploying a mass spectrometer is proposed for China future lunar exploration missions. In order to receive the lunar regolith sample from a robotic arm with a soil sampler and transfer it into a furnace for further analysis, a sample manipulation mechanism is required during the above work flow. To solve the problems of adapting the sampler's docking accuracy, receiving and transferring two different types of lunar soil sample under times of in-situ analysis, etc., a sample repetitive manipulation mechanism (SRMM) is proposed in this paper. By using a floating adjustable docking components and a flexible hopper, two types of encapsulated regolith sample and bulk material sample can be received with minimal sample loss, respectively. In order to receive and transfer two types of samples multiple times, two sample receiving methods have been designed that can be repeatedly transferred. A worm and worm wheel combined with a ball screw is designed in SRMM. To verify the above mechanism design, validation experiments were conducted. It indicates that this novel SRMM can be deployed in the future mission after further environmental tests.

Keywords: New theory and technology in robotics and biomimetics
Abstract: Human bones have formed the preferred cofiguration for high-strength and lightweight after long-time evolution. Taking human’s longest and strongest bone - the femur - as an example, it is consist of two characteristic layers, i.e. the substantia compacta and the substantia spongiosd. This article innovatively imitates the structural characteristics of human femur, the thigh of humanoid robot is designed in form of "variable thickness shell+variable density lattice". The thickness of shell and the density of lattice are adjusted by the initial stress distribution individually. Results show that the weight of shell and lattice of the thigh structure can be reduced by 20% under reasonable mapping relationship of "stress - shell thickness" and "stress - lattice rod diameter", while the structural stiffness meets the application requirements. Finally, the limiting factors of the "variable thickness shell+variable density lattice" structure designing approach are analyzed, and potential measures for optimizing the design method of the humanoid robot thigh in the future are described.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics
Abstract: In recent years, biomimetic robotics has emerged as a promising field that draws inspiration from nature to develop innovative and efficient robots. In this study, we have designed a biorobot capable of crawling and barrier-crossing, inspired by the walking gaits of crabs. Firstly, we analyzed the motion of an individual crab leg to determine the required degrees of freedom for crawling and barrier-crossing. Then, we utilize the screw theory to synthesize the mechanism of a single branch. Finally, we select a suitable parallel mechanism configuration for this research. Simulation analysis is conducted to test the variations in actuation at different driving points under various pose states. We observe stability in changes in driving points with respect to the retraction angle of the leg, indicating excellent obstacle overcoming capabilities possessed by this robot.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Space robots, aerial robots, and underwater robots, New theory and technology in robotics and biomimetics
Abstract: Water-jet propulsion helps flying squid shoot quickly from underwater into the air, overcoming the huge differences in density and viscosity between two fluids. To get faster underwater takeoff, water-jet propulsion can be great inspiration for the design of new propulsion system for aquatic unmanned aerial vehicle (AquaUAV). In this paper,an novel aerial-aquatic water-jet thruster is developed and fabricated, which uses butane and oxygen to realize explosive water-jet. An unit butane cell and oxygen cell can satisfy more than 50 explosions without replacement of components.We designed different fine-tuning strategies for the two gas supply, designed and completed experiments to obtain stable gas supply and controllable explosion. And the optimized supply systems were tested on the force measurement system. For the first time, our new thruster achieved underwater no-exchange-air explosion, which may help AquaUAV handle continuous underwater explosive propulsion and repeatable launches other than propellers.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics
Abstract: This paper presents the design of an underwater snake-like robot featuring four fully submersible joints and verifying its primary underwater properties. The actuators that drive these joints are designed to focus on minimal friction, have a lubricant-free gear reducer, and have no waterproof sealing, making them suitable for direct exposure to underwater experiments. The paper thoroughly discusses the efficiency and rated torque measurement of these actuators, the design of the snake robot, and the primary underwater verification. The ultimate objective of this snake robot is to enable it to engage in interactive tasks that require direct mechanical interaction with the underwater surroundings. The execution of various tasks with this snake robot remains part of future work. The robot demonstrates serpentine and eel-like locomotion in the underwater environment, showcasing its adaptability and potential for various applications in underwater environment.

Keywords: New theory and technology in robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Smart sensors and actuators
Abstract: In order to improve the overall deflection range of a multi-posture actuator (MPA) based on shape memory alloy (SMA) and reduce SMA usage, a 3D printed optimized VT-type SMA actuation module (VSM) is proposed for the MPA in this paper. Specifically, the VT structure parameters and the length of SMA wire in single module are set as the optimization variables, and the maximum deflection angle and the shortest length of the SMA wire in VSM are used as the optimization goals. The nondominated sorting genetic algorithm II (NSGA-II) is used to obtain the Pareto optimization frontier solution set. Then a set of the optimized Pareto solutions are selected for simulation and experimental verification. The results show that the optimized VSM is 511% and 456% higher than the initial design angle deflection in simulation and experiment, respectively, and the amount of SMA wire used is reduced by 10%. In addition, at the expense of SMA wire usage, the actuation force can be increased by adding the number of parallel SMA wires. This work lays the foundation for future research on modular miniaturized MPAs based on SMA actuation.

Keywords: Soft robotics and liquid-metal robotics, Smart sensors and actuators, Applications of robotics and biomimetics
Abstract: In this study, we propose a soft diaphragm pump driven by circular dielectric elastomer (DE) as the active diaphragm coupled with magnetic attraction as nonlinear antagonistic mechanism to improve amplitude. By taking the principle of volume change in diaphragm chamber together with two check-valves on both sides, directional pumping is achieved. Despite the presence of damping in the passive valves, the flexible pump still exhibits a good output performance at the resonance of DE. Furthermore, by optimizing the distance between the magnets and increasing the applied electric field strength, the output performance of the flexible pump can reach a maximum output pressure of 28 mbar and a peak gas flow rate of 0.24 L/min at an applied electric field of 60 MV/m.

Keywords: New theory and technology in robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics
Abstract: Hyper-redundant robot is a type of special robot often used for detection and maintenance work in confined environments. With continuous research in recent years, various types of hyper-redundant robots have emerged. However, the dynamic modeling of hyper-redundant robots remains challenging. To address this issue, this paper introduces a hybrid degree-of-freedom configuration of hyper-redundant robots (HDHR), derives its kinematic model, and uses a modular dynamic modeling method to construct the dynamic model of hyper-redundant robot. Simulation results verify the feasibility of dynamic mode.

Keywords: Smart sensors and actuators, Robotics in intelligent manufacturing, Applications of robotics and biomimetics
Abstract: For solving the problem of low efficiency and poor accuracy of adjacent pipe interface alignment when laying large-bore pipelines, a method of relative pose detection of adjacent pipe interfaces based on the point laser ranging sensors was proposed. The method adopts the data from point laser ranging sensors to fit an ellipse of the pipe to be aligned and compare the ellipse with the cross-sectional circle to obtain the relative pose of two adjacent pipe interfaces. The analytical model of the relative pose between the ellipse axis of the pipe to be measured and the cross-sectional circle axis was constructed, and a pipe interface relative pose detection device was designed and built. Finally, an experiment of measuring the angle between axis vectors of two pipes was designed to verify the accuracy and robustness of the proposed method. The experimental results indicated that the measurement error of the angle between axis vectors of two pipes is 0.3°, which can meet the requirements of the standard of large pipe interface alignment operation.

Keywords: Space robots, aerial robots, and underwater robots, New theory and technology in robotics and biomimetics
Abstract: In the era of Industrial 4.0, power safety is essential to industrial production. The trees growing around the transmission network threaten the security of electric power, while the traditional cleaning methods are inefficient and provide poor security. In this paper, a kind of aerial robot that can be used to clean up trees is designed and implemented. For tree-cleaning, the dynamic model of the robot and the unified mechanical model of contact with trees are established based on the extended Lagrangian method. Then, the wrench observer is designed through this mechanical model, and the robot’s control system is created by combining the observer with the auto-disturbance rejection control method. The experimental results show the method is feasible and can efficiently carry out the tree-cleaning operation.

Keywords: Artificial intelligence in robotics, Human-robot interaction, Medical robotics, biomedical and rehabilitation engineering
Abstract: For unilateral pathologies, effective rehabilitation relies on the use of a customised trajectory in order for the user to relearn a natural and symmetrical gait. In recent years, lower-limb exoskeletons have seen a growing interest due to their capacity to provide support and facilitate repetitive exercises while correcting the user’s motion. However, in the context of robotic-assisted locomotion, the investigated trajectory models tend to rely on generating standardised walking patterns that lack step-specific customisation, and therefore do not account for the dynamic variations of natural gait. This paper investigates the viability of an echo-based approach for trajectory generation, which centres around the dynamic relabelling of a time-invariant reference trajectory, based on the motion of the contralateral leg. The presented cascaded network combines (1) a classifier that determines the gait phase performed by the sound leg and updates the reference trajectory accordingly, with (2) a regressor that uses electromyography inputs from the investigated leg to predict the gait cycle percentage performed, and provide the associated knee angle based on the dynamic reference. This trajectory generation framework was evaluated on 6 able-bodied subjects, using both steady-state and transient speeds. Despite some discrepancies in the range of motion, the produced knee angle trajectory strongly resembles the experimentally captured ones for both conditions, with an average mapping Root Mean Squared Error across subjects of 4.62◦±0.39◦ for steady-state and 5.88◦±1.83◦ for transient speeds. This proof-of-concept implementation demonstrates the potential of an echo-based approach for personalised dynamic trajectory generation in unilateral exoskeleton ap

Keywords: Medical robotics, biomedical and rehabilitation engineering, New theory and technology in robotics and biomimetics, Smart sensors and actuators
Abstract: Assistive exoskeletons are medical devices typically designed to perform repetitive movements. Recent researches highlighted the necessity to endow such devices with perceptual capabilities, in order to perform adaptive movements. This would allow using exoskeletons in everyday activities such as, for example, the ascent of a staircase. This work addresses some of the typical problems related to the motion along a staircase of a lower limb exoskeleton. In particular, the focus will be posed on problems such as the identification of the staircase and the relative position between them and the exoskeleton, the motion planning for the ascending foot and its adaptation to the changes of the torso orientation. The proposed strategies have been experimentally verified and the results are documented through videos.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Soft robotics and liquid-metal robotics, Smart sensors and actuators
Abstract: Wearable devices take essential support to help elders or disabled individuals. State-of-art walking-assistance devices usually utilize the gravitational potential energy to provide continuous powers for wearable actuators. Here in this work, we proposed a novel design of a self-powered and wearable pneumatic artificial muscle device. When walking or running, the plantar pressure provides quasi-periodic energy to compress the airbags embedded within the shoe pads, thus to generate on-demand compress air for pneumatic artificial muscles tied to the calf. To set up a self-powered insole actuator, the quasi-periodic plantar pressure energy is recycled and regulated by a pressure control system. The geometry of airbag unit is investigated to provide optimized supporting forces. With a 2×2 airbag array, the self-powered insole actuator is well fitted to daily-life shoes, and it could provide 30 N supporting forces at the calf artificial muscle terminal. It was experimentally demonstrated that the self-powered pneumatic artificial muscle was able to meet the requirements of walking assistance, without the bulky compressed air supply equipment.

Keywords: Human-robot interaction, Medical robotics, biomedical and rehabilitation engineering
Abstract: This paper presents a novel approach to measuring upper limb muscular manipulability considering human biomechanics. We address the limitations of classical manipulability measures in robotics when applied to the human body. Our method introduces the concept of a force envelope to estimate the capability of the human arm to exert forces in different directions, considering the contributions of the muscles. To achieve this, we employed a biomechanical model based on Hill’s muscle model, calibrated using both geometric (segmental lengths) and strength-based (muscle activation) approaches to adapt to individual users. Furthermore, we designed a control algorithm that enables a robotic device to assist the user in unfavorable directions, guided by the manipulability measure. By providing a more isotropic response, the robotic device compensates for low manipulability in certain regions of the workspace. We conducted experiments using a haptic robot in admittance mode along the sagittal plane, where a viscous environment acted as a load to hinder human movement throughout the workspace. Our results demonstrate the effectiveness of the proposed method in reducing human effort by assisting in less manipulable directions while leaving high manipulability directions unassisted. Additionally, we successfully verified the superiority in performance of our novel approach against existing methods.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Human-robot interaction, Artificial intelligence in robotics
Abstract: Brain-computer interfaces (BCIs) have emerged as a groundbreaking technology enabling direct communication between the human brain and external devices. Yet, the attainment of higher accuracy rates while making classifications remains a problem and demands the development of more sophisticated machine learning architectures and the exploration of novel feature domains. To address these issues, we have presented a novel approach for enhanced EEG-fNIRS (electroencephalography-functional near-infrared spectroscopy) classification through a concatenated convolutional neural network (CNN) with band analysis. Our proposed methodology leverages the power spectral density (PSD) of EEG signals in specific frequency bands and captures metabolic changes in the brain through fNIRS data. Then it employs a concatenated CNN architecture for improved feature extraction and classification in EEG-fNIRS data. Through an experimental evaluation on a public dataset containing three mental workload (MWL) tasks, our proposed model demonstrates superiority compared to previous studies by achieving higher accuracy rates of 93.8%, 97.2%, and 94.5% for n-back, discrimination/selection response (DSR), and word generation (WG) tasks, respectively. Furthermore, our results emphasize the advantages of utilizing a hybrid EEG-fNIRS approach over unimodal EEG or fNIRS methods. These outcomes underscore the potential of our multimodal EEG-fNIRS approach and its applicability in BCI, neurorehabilitation, and human-computer interaction contexts.

Keywords: Soft robotics and liquid-metal robotics, Applications of robotics and biomimetics
Abstract: This paper introduces a full system modeling strategy for a syringe pump and soft pneumatic actuators(SPAs). The soft actuator is conceptualized as a beam structure, utilizing a second-order bending model. The equation of natural frequency is derived from Euler's bending theory, while the damping ratio is estimated by fitting step responses of soft pneumatic actuators. Evaluation of model uncertainty underscores the robustness of our modeling methodology. To validate our approach, we deploy it across four prototypes varying in dimensional parameters. Furthermore, a syringe pump is designed to drive the actuator, and a pressure model is proposed to construct a full system model. By employing this full system model, the Linear-Quadratic Regulator (LQR) controller is implemented to control the soft actuator, achieving high-speed responses and high accuracy in both step response and square wave function response tests. Both the modeling method and the LQR controller are thoroughly evaluated through experiments. Lastly, a gripper, consisting of two actuators with a feedback controller, demonstrates stable grasping of delicate objects with a significantly higher success rate.

Keywords: Artificial intelligence in robotics, Human-robot interaction, New theory and technology in robotics and biomimetics
Abstract: In recent years, reinforcement learning has achieved significant advances in practical domains such as robotics. However, conveying intricate objectives to agents in reinforcement learning (RL) remains challenging, often necessitating detailed reward function design. In this study, we introduce an innovative approach, MEETRE, which integrates max-entropy exploration strategies with random encoders. This offers a streamlined and efficient solution for human-involved preference-based RL without the need for meticulously designed reward functions. Furthermore, MEETRE sidesteps the need for additional models or representation learning, leveraging the power of randomly initialized encoders for effective exploration.

Keywords: Artificial intelligence in robotics, Smart sensors and actuators, Robotics in intelligent manufacturing
Abstract: Kinesthetic demonstration requires accurate force/torque measurements from sensors. However, even in static conditions, the sensor readings exhibit non-zero fluctua- tions, which can lead to unstable force control and impair the quality of kinesthetic demonstration. In this paper, we refer to the above problem as phantom zeros and systematically analyze factors contributing to their variability of a traction force sensor. Neural networks (NN) are first introduced to model the complex nonlinear mapping between sensor orientations and phantom zeros. The initial parameters of the NN are then optimized using a genetic algorithm (GA) to prevent convergence to local optima and improve modeling accuracy. In addition, we develop an experimental platform with a physical UR10 robot and a custom traction sensor to comprehensively evaluate the proposed approach. Results demonstrate that the GA-optimized NN achieves higher precision and robustness in predicting phantom zeros under different orientations com- pared to least squares and vanilla NN baselines. By modeling and predicting phantom zeros, the proposed method can filter out phantom force fluctuations during kinesthetic demonstra- tion, while preserving critical motion information. This work provides insights into modeling and mitigating force/torque sensor uncertainties for enabling more precise robot control and interactive guidance.

Keywords: Autonomous mobile robots and manipulators
Abstract: Iterative linear quadratic regulator (iLQR) has gained wide popularity in addressing trajectory optimization problems with nonlinear system models. However, as a model-based shooting method, it relies heavily on an accurate system model to update the optimal control actions and the trajectory determined with forward integration, thus becoming vulnerable to inevitable model inaccuracies. Recently, substantial research efforts in learning-based methods for optimal control problems have been progressing significantly in addressing unknown system models, particularly when the system has complex interactions with the environment. Yet an incremental machine learning method or a deep neural network is normally required to fit a substantial scale of sampling motion data. Existing works mainly focus on capturing the whole picture of the motion pattern from trial data (which is essentially noisy and hard to identify with any simple predefined structures) for complex robot systems. Instead, we present Neural-iLQR, a learning-aided shooting method over the unconstrained control space, in which a neural network with a simple structure is used to drive the optimization process towards optimal direction. Through comprehensive evaluations on two illustrative control tasks using three different kinds of lightweight network structures, the proposed method is shown to outperform the conventional iLQR significantly with fast and continuous cost convergence in the presence of inaccuracies in system models, which demonstrates the generalizability and robustness of the proposed learning-aided method.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Artificial intelligence in robotics, New theory and technology in robotics and biomimetics
Abstract: Bipedal humanoid robot has the ability to both move and manipulate in complex environments, which is of great significance in the future. However，stable bipedal walking in the real world has always been a challenge in industry and even in academia. The traditional model-based methods are highly dependent on the environment, with high modeling complexity and lack of generalization. The solution based on the simplified model usually causes the problem that the control algorithms cannot adapt to complex terrain environment. This paper presents a newly designed bipedal humanoid robot, Xiao-Man. Aiming at achieving the robot’s terrain-adaptive walking behavior, a reinforcement learning based Actor-Critic network with asymmetric structure is proposed. Without using any external perception information, robust bipedal walking behavior of Xiao-Man is achieved. In the process, we also build the dataset based on the joint actuation truth data and train a joint actuator network to reduce the gap between the expected torque and the actual response torque. Experimental results show that the bipedal humanoid robot equipped with the trained control policy achieves the capability of stable walking and disturbance rejection only rely on proprioceptive information.

Keywords: Soft robotics and liquid-metal robotics, Applications of robotics and biomimetics
Abstract: Soft robots have shown their value as alternatives or supplements to rigid robots in applications like search and rescue missions and complex precise tasks. Their ability to take on various shapes and apply adaptable force gives them an advantage over stiff robots. However, sometimes their soft structure doesn't offer enough force for the task. Hybrid soft robots (HSRs) combine a soft body with a stronger backbone to handle tasks needing more strength. This rigid part lets us use rigid body dynamics to estimate HSR behavior. Here, we introduce a simplified N-link rigid body dynamic model with constant stiffness to mimic HSR behavior. While soft robots' stiffness varies, the backbone in HSRs makes it similar to having constant stiffness. Comparing experiments supports the effectiveness of our N-link model for HSR modeling.

Keywords: Applications of robotics and biomimetics, Medical robotics, biomedical and rehabilitation engineering, Robotic vision and image processing
Abstract: Soft robots are a hot spot in today's robotic research, because most of them exist in the form of continuums, and the current continuum is difficult to recognize the shape and reproduce the corresponding shape. In this paper, we propose a method, in which the shape features of the flexible continuum are obtained by contour centerline extraction and binocular camera reconstruction and the modeling of the relationship between the motor input and the shape output of the continuum is completed using neural networks. Simulation environment is set up to test the shape estimation and shape control of the flexible continuum. Results show that this method can prediction and reproduce the shape of the continuum well. This method can be used in shape control of the continuum robot.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Soft robotics and liquid-metal robotics, New theory and technology in robotics and biomimetics
Abstract: Taking inspiration from the two turning mechanisms exhibited by biological snakes, this paper presents two turning methods for soft underwater snake robots: head steering and unbalanced gait steering. The structural design, driving principles, and turning motion model of the proposed snake are introduced. Through experimental validations under the same conditions, the results demonstrate the turning feasibility of two steering methods in underwater snake robots, revealing superior turning capabilities of head steering compared to unbalanced gait steering utilized by conventional ground-based snake robots. This paper pioneers head-turning investigation in underwater snake robots and offers insights for future applications involving the agile maneuvering of underwater robots in aquatic environments.

Keywords: Soft robotics and liquid-metal robotics, Applications of robotics and biomimetics
Abstract: In the realm of soft robotics, the pneumatic system is a crucial component that plays an essential role in achieving optimal motion performance since all motion performance is ultimately linked to precise pressure control within the air chambers. To efficiently provide programmable pressure states for driving soft robots, this paper proposes a programmable pressure pneumatic system consisting of air pumps and valves with eight independent air channels. This system can provide precise hybrid negative and positive pressure for soft pneumatic robots. Closed-loop pressure control is achieved by sensing the pressure states in each air channel through sensors. Furthermore, two types of soft robots based on single-chamber soft pneumatic actuators are designed to verify the effectiveness and practicality of the pneumatic system. The proposed pneumatic system is a universal platform that can meet the pressure control requirements of various soft robots driven by both positive and negative pressure and holds significant implications for actuation and control in soft robotics.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Soft robotics and liquid-metal robotics, Applications of robotics and biomimetics
Abstract: This paper proposes a snake-like robot with two soft actuator modules in series, and its gait strategies are studied. Firstly, a bellows-type soft actuator was designed and then manufactured by the paraffin loss method. Secondly, the driving system is designed according to the fluid driving principle of the soft actuator, and the setting of critical parameters is studied. Finally, the structure of a soft snake-like robot is demonstrated, and two gait strategies are proposed. Experiments are carried out to test the motion speed of the robot when different gait strategies are applied. The results show that the soft snake-like robot achieves a speed of 11 mm/s (0.03 body length/s) when using the phase offset strategy without delay activation. This shows its application potential in search, detection, rescue, and other operations in a narrow environment.

Keywords: Soft robotics and liquid-metal robotics, Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics
Abstract: In this work we present the design and development of a twitching control system for pneumatic artificial muscle (PAM) actuators. In this system, a shape memory alloy (SMA) poppet valve is used to control the flow of air pressurizing the inner bladder of a McKibben-based PAM. The valve is driven by pulse width modulation with changing frequency, allowing the PAM to resemble the mammalian muscles twitching. An adjustable passive exhaust orifice is used in place of an exhaust air valve to release the air pressure from the bladder mimicking natural muscle relaxation behaviour. Experiments under isometric and isotonic contractions show that the proposed control system reduces hysteresis of about 72% compared to continuous activation against 15% loss in contraction force with respect to conventional exhaust air valves.

Keywords: New theory and technology in robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Smart sensors and actuators
Abstract: Most control methods for soft robots are developed based on kinematic models derived from deformation and force analysis. However, due to the non-linearity and uncertainty of soft robotic structure, it is difficult to establish accurate models, leaving a great gap in the precise control of soft robots. Recent research has shown that machine learning provides a highly effective solution. In this work, we propose a back-propagation neural network based on particle swarm optimization algorithm to establish the surrogate kinematics of a airbag-type soft actuator. Using the motion data of the soft actuator to train the network model, the corresponding relationship between the soft actuator and the end position can be obtained. The results show that the surrogate model has a good prediction effect, and the average relative error of the model is 6.4%, enabling control the motion of the soft actuator accurately enough.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Soft robotics and liquid-metal robotics, Human-robot interaction
Abstract: This paper focuses on the design and systematic evaluation of fabric-based, bellow-type soft pneumatic actuators to assist with flexion and extension of the elbow, intended for use in infant wearable devices. Initially, the performance of a range of actuator variants was explored via simulation. The actuator variants were parameterized based on the shape, number, and size of the cells present. Subsequently, viable actuator variants identified from the simulations were fabricated and underwent further testing on a physical model based on an infant's body anthropometrics. The performance of these variants was evaluated based on kinematic analyses using metrics including movement smoothness, path length, and elbow joint angle. Internal pressure of the actuators was also attained. Taken together, results reported herein provide valuable insights about the suitability of several actuator designs to serve as components for pediatric wearable assistive devices.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Soft robotics and liquid-metal robotics, New theory and technology in robotics and biomimetics
Abstract: Soft Magnetic Manipulators (SMM) have the potential to improve current endoscopic procedures through extreme miniaturization and inherently safe tissue interaction. This relatively new field of soft robots is being extensively researched for various medical applications. However, utilization of two or more SMMs under stable control within the same workspace is challenging due to their exposure to similar or identical applied magnetic fields and field gradients. To improve stable independent control of SMMs in a shared workspce, we explore geometric varations to the SMM design, consisting of elastomeric double helices wrapped around a cylindrical core. The influence of geometrical properties on stiffness in two bending planes is investigated, as well as torsional stability across six tested designs. Experimentally acquired data on bending and twisting of proposed designs is compared to equivalent soft magnetic cylinders. Finally to demonstrate the capability of independent actuation, two selected designs are tested in multi-SMM parallel configurations, consisting of two and three parallel SMMs within the same workspace.

Keywords: Soft robotics and liquid-metal robotics, Human-robot interaction, New theory and technology in robotics and biomimetics
Abstract: The Piecewise Constant Curvature (PCC) assumption is the most widely used in modeling and control of soft robots. However, a limitation of PCC models lies in accurately describing the deformation of a soft robot when executing dynamic tasks such as operating under gravity or interacting with humans. This paper introduces a new methodology for dynamic modeling and control of a multi-segment soft arm in the 3D space where each segment undergoes non-constant curvature deformations. In this framework, the soft manipulator is modeled as a series of segments, with each one represented by two stretchable links connected by a universal joint. Furthermore, we devise and analyze a controller for motion control in the configuration space with unknown load. The controller is implemented on a four-segment soft robotic manipulator and validated in a range of dynamic trajectory tracking tasks.

Keywords: Medical robotics, biomedical and rehabilitation engineering, New theory and technology in robotics and biomimetics
Abstract: Continuum robots have the great potential in endoluminal diagnosis and intervention toward narrow and confined workspace. However, fabrication of continuum robots still poses a grant challenge in its miniaturization. To solve this, this paper designs a thermally drawn monolithic polymer continuum robot with laser cutting profile. First, the continuum robot is designed with two degrees of freedom (DOF) for shape deflection, and it consists of four optical fibers with the fiber Bragg grating (FBG) sensor on each for tension sensing as driving cables. Second, its main body is a multi-channel polycarbonate (PC) tube produced by thermal drawing process, and a series of rectangular cutouts are made on it via laser cutting process. By combining these two fabrication processes, the developed continuum robot realizes the integration of the multi-channel structure and the notch feature. Third, a control framework of puller-follower controller using the integrated tension sensing is proposed for teleoperation control. Finally, experimental setup is built for the robot validation, and results show that the repeated positioning accuracy of the robot is 1.08 mm. Moreover, a micro endoscope and a needle are added to the end of the continuum robot, and a phantom study of auripuncture and drug delivery is completed to verify the preliminary feasibility. And the proposed fabrication approach enables the miniaturization and complex structure for the cable-driven continuum robots.

Keywords: Soft robotics and liquid-metal robotics, Smart sensors and actuators, New theory and technology in robotics and biomimetics
Abstract: Soft actuator plays a critical role in soft robotics. Mechanical characterization of the soft materials serves as a must-do process for accurate modelling, simulation and control of soft actuators. Uniaxial tensile test, as a fundamental testing modality, of fragile biological materials like hydrogel used in soft robotics, biomechanical systems and tissue engineering remains a challenge, because of their poor mechanical properties (kPa ~ MPa), sample gripping difficulties, low force range and hence inaccuracy of force sensing caused by friction of mechanical guidance. In this paper, a linear soft actuator with magnetic fixture and actuation is developed to address the mentioned problems. This device offers a loading range of 0.1 mN to 0.1 N and a corresponding stretching range of a few millimeters. A USB-microscope is employed for remote detection of displacement and force based on image processing and a magnetic-simulation-based lookup table. Maxwell model of viscoelastic material is adopted for system modelling and a closed-loop PI controller is designed here for steady force and displacement outputs. Finally, elastic, creep and relaxation tests are implemented on hydrogel and silicone materials. The reasonable results prove the feasibility of this measurement approach. This device stands out among various mechanical testing equipments due to its contactless, friction-free, low-cost and accurate operational modalities.

Keywords: Smart sensors and actuators, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Space robots, aerial robots, and underwater robots
Abstract: Soft underwater robots have attracted widespread attention from researchers due to their excellent lightweight, low noise actuation, and excellent environmental interaction capabilities. A potential challenge is to develop a suitable flexible driving method for underwater robots using soft materials, which can not only successfully replicate the driving characteristics of biological muscles, but also replicate the swimming characteristics of aquatic organisms as much as possible. Here, through bionic the deformation feature of fish muscles, an underwater flexible actuating method based on dielectric elastomer actuators (DEAs) has been proposed, which can realize the large bending amplitude of the robotic fish caudal fin. By controlling the two driving units of a single soft actuator, the periodic swinging motion of the caudal fin can be simulated. In addition, the oscillation characteristics of the robotic fish caudal fin are simulated and analyzed by COMSOL software The results show that the maximum unidirectional bending angle of the robotic fishtail can reach 20° and the simulation results show that alternating anti-Carmen vortex structures can be generated, which proves the effectiveness of the proposed flexible actuating method. Finally, a robotic fish design concept composed of 3D-printed flexible shells and two soft actuators has been proposed, which can realize continuous C-shaped and S-shaped body curves by controlling the amplitude and phase of the soft actuator. This study not only provides a suitable driving method for the design and development of soft underwater robots, but also has guiding significance for the design of other soft actuators, dexterous operations, and soft robots.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: High-precision dynamic model is fundamental for achieving excellent effect in path following control of snake robots. In the field of control, data are increasingly used to model dynamical systems, as it allows the direct assessment of residual model uncertainty. This paper proposes a novel model predictive path following control architecture based on a physics-data hybrid model for planar snake robots. The hybrid model integrates a physics based first principle dynamics model with an additive Gaussian process (GP) regression based residual model part. Considering the control input constraints problem, the model predictive control (MPC) method is adopted to realize the path following with velocity tracking of the snake robot. GP regression is a powerful tool for model learning due to its flexibility and inherent ability to describe uncertainty in function estimation. However, when GP regression is used to predict the residual model and the control architecture combined with MPC, it is necessary to consider the features selection according to the characteristics of the snake robot and the balance between control effect and computational complexity. Aiming at these problems, specific solutions are given in this paper. Finally, simulations for the planar snake robot are conducted to showing the performance improvements of the proposed control design compared to the typical snake robot control method.

Keywords: Autonomous mobile robots and manipulators, Applications of robotics and biomimetics, Artificial intelligence in robotics
Abstract: In light of past epidemics, people have become more aware of the possibility of viruses on object surfaces and in the air. As a result, there is an increasing demand for robots capable of autonomously disinfecting surfaces and purifying the air. This paper focuses on a mobile robot equipped with air purification and UV-C LED disinfection to combat airborne pathogens and environmental pollution. Through the onboard sensors, the proposed algorithm uses gaussian process model to predict the high potential air pollution area in real-time without the requirement of a distributed array of sensors, and it avoids environmental modifications. Then we propose the hybrid artificial potential field (HAPF) by integrating the predicted result with the artificial potential field method, to generate a heuristic path that can navigate the robot to the affected areas with obstacle avoidance. Experimental testing in simulated and real-world environments demonstrates the effectiveness of the proposed approach in rapidly planning purification paths, marking significant progress in automated air purification, and disinfection technology.

Keywords: Human-robot interaction, Autonomous mobile robots and manipulators
Abstract: Recently, many researchers propose different methods to solve the motion planning problem in the human-robot co-existing environments. However, when encountering crowds that suddenly appear, conventional motion planning algorithms often generate similar homotopy paths with slight local variations, which may not generate new feasible paths in time and thus lead to congestion for the robot. In order to address this problem, this paper proposes a knowledge-based full-time non-homotopy path optimization method through online environmental learning, which generates heuristic paths to guide motion planning by combing the pedestrian knowledge and the environmental structure. Firstly, a pedestrian matrix based knowledge base is proposed to record the pedestrian flow pattern with respect to different locations, and the corresponding pedestrian matrix is extracted according to the perception during the robot navigation process. Secondly, the environmental perception is performed surrounding the key obstacles that appears around the crowds to extract the feature points. Then, the extracted feature points are reused as prior knowledge in subsequent navigation. When crowds suddenly appear in the global path, the algorithm will prioritize searching a non-homotopy path to avoid crowds. The experimental results show that the method proposed in this paper can significantly reduce the time required for the robot to navigate and improve the success rate of navigation in crowded environments.

Keywords: Artificial intelligence in robotics, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: The humanoid gait generation task of biped robots has been a lasting challenge. To achieve a stable gait in such a complex kinematic system, machine learning methods such as reinforcement learning (RL) are widely applied to this topic and made progressive results. RL provides a convenient end-to-end solution and has become the mainstream machine learning method in the field of robotics. However, there are still problems when using traditional RL methods in aspects of similarity to human natural gait, training time, robustness to perturbances, and generalization ability of other tasks. In this paper, we propose an improved framework based on the feedforward enhanced reinforcement learning (FERL) algorithm to efficiently generate a humanoid gait for a small humanoid robot KRH-3HV. In FERL, the control action of biped robots consists of two parts: the reference part and the weighted RL part. In this paper, we introduced prior knowledge that human walking gait exhibits sinusoidal characteristics and designed reference actions by inverse kinematic analysis based on this principle. The weighted RL part is independently set to motivate the biped robot to complete the target task. Providing a reference control signal sequence helps to generate an ideal gait and the action space of the weighted RL is decreased compared with traditional RL so that the training time is reduced while still maintaining certain robustness. By setting up 4 groups of control experiments in a simulation environment, we efficiently generated an effective humanoid gait to complete the target tasks of walking on flat ground and climbing stairs, and proved that the proposed method works well in gait biomimetic similarity, training efficiency, robustness, and generalization ability to other tasks.

Keywords: Autonomous mobile robots and manipulators
Abstract: Due to the widespread use of robotic arms, path planning for them has always been a hot research topic. However, traditional path planning algorithms struggle to ensure low disparity in each path, making them unsuitable for operation scenarios with high safety requirements, such as the undercarriage environment of train. A Reinforcement Learning (RL) framework is proposed in this article to address this challenge. The Proximal Policy Optimization (PPO) algorithm has been enhanced, resulting in a variant referred to as Randomized PPO (RPPO), which demonstrates slightly accelerated convergence speed. Additionally, a reward model is proposed to assist the agent in escaping local optima. For modeling application environment, lidar is employed for obtaining obstacle point cloud information, which is then transformed into an octree grid map for maneuvering the robotic arm to avoid obstacles. According to the experimental results, the paths planned by our system are superior to those of RRT* in terms of both average length and standard deviation, and RPPO exhibits better convergence speed and path standard deviation compared to PPO.

Keywords: Smart sensors and actuators
Abstract: Multi-dimensional force sensors commonly utilize sensing beam structures for pursuing their high sensitivity and multi-axis measurements. To achieving high-performance of such sensors, research has predominantly focused on material improvements, sensor configurations, and efficient algorithms. However, leaving alone the sensor beam, one of the core components, relatively unexplored amid rapid technological advancement. This study addresses this gap by investigating into sensor beam designs and refinement, aiming to enhance sensor performance. The potential for improved accuracy, stability, and adaptability through the sensor beam refinement underscores the significance of this research. In detail, the study commences with an examination of strain distribution across sensor beams of varying geometries, facilitating the identification of an optimal configuration tailored to the project's application. In addition, the slot size adjustments are applied to the beam design, yielding improvements in sensor performance. Conclusively, an analysis of frequency response characteristics is conducted to ascertain the superior design iteration for the designed sensor beams. This investigation underscores the paramount importance of beam design refinement in augmenting multi-dimensional sensor functionality.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: The locomotion of snakes is unique as it allows snakes to efficiently navigate complex and uneven terrains without articulated limbs. This movement pattern is attractive in the field of robotics exactly for its hyper redundancy, versatility, and lack of requirement for relatively complex locomotive systems. Snake-like robots that imitate these movement patterns are uniquely suited for use in extreme and hostile environments like deep sea exploration and outer space. Controlling robots in environments like these requires lots of training and intense concentration during operation. Voice User Interfaces (VUIs) can be used to bypass much of the need for this training and attention to operation by abstracting the direct control of a robotic system into the domain of speech. The design and implementation of a VUI demonstrating this idea is discussed here. The basic structure of a VUI is described and the implementation of a VUI in conjunction with a snake robot is shown. In experimentation the performance of the VUIs components were evaluated and the navigation of an obstacle course via the developed VUI and a remote controller were compared. The command recognition capability of the VUI was found to be 96% and the ability of the VUI to enable navigation of the obstacle course was found to be comparable considering the differences of the control formats.

Keywords: Autonomous mobile robots and manipulators, Robotic vision and image processing, Robotics in intelligent manufacturing
Abstract: Ball-on-plate balancing is a popular challenge for the implementation of control systems. In this paper, the Universal Robots UR3 industrial arm is set up for the challenge. An acrylic plate is firmly grasped by the robot’s gripper. The pure rotation about three axes is achieved at the center of the plate by considering the offset of the plate from the tool frame. A web camera is mounted on the last link to observe the position of a ping pong ball moving on top of the plate. The color-based object detection is used with OpenCV and the coordinates of the ball are mapped to the pixel of the camera. At the beginning of each test, the ball was placed away from the center of the plate. Considering the x-axis and y-axis error of the ball position away from the center of the plate, the robot arm stabilizes the ball to the center according to the PID controller gain configurations. The effect of the proportional gain on reducing the rise time of the response is observed. The advantage of the derivative gain on reducing the response overshoot is observed. The advantage of the integral gain on reducing the steady-state error is not clearly observed for this ball-on-plate balancing problem. The implemented controller is robust against the small vibration of the camera with respect to the robot arm.

Keywords: Autonomous mobile robots and manipulators, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics
Abstract: Motion planning for quadrupedal robots on unstructured terrains demands the consideration of torso terrain adaptation and safety foothold placement. This paper presents an autonomous and safety stair climbing control strategy, aiming to simultaneously optimize body and leg movements. An improved mobility metric taking into account leg mobility and maintaining stability margin, is introduced. By penalizing deviations from the current joint angle and nominal leg configuration in the cost function, the stance legs can naturally extend to lifting body, and the projection of the center of mass is always inside the support region. Then, the reference climbing velocity is automatically determined based on the geometric information of the stairs. With these reference velocities, we solve a two-layer control framework which couples nonlinear model predictive control (MPC) with whole-body control (WBC), and add several different control barrier functions (CBF) constraints to ensure safety foot placement and edge avoidance. Note that each constrained act on the kinematics or dynamics level depends on the order of the system dynamics. In the simulation, we demonstrate that the robot efficiently climbs several different size stairs with 10 cm height (25% of the maximum leg length) and the maximum climbing velocity reaches 0.64 m/s.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: In the field of bipedal locomotion control, the Linear Inverted Pendulum Model (LIPM) is widely applied to approximate the walking dynamics of bipedal robots. The linear velocity of the center of mass serves as a controlled state variable, enabling real-time high-dynamic control. This paper introduces a solution that adopts a lateral segmented structure and a telescopic leg design for a five-degree-of-freedom underactuated bipedal model, aiming to reduce the number of motors and the overall mass of the robot. A LIPM based on angular momentum prediction is devised, amalgamating foot position prediction with angular momentum. Due to the inherent inaccuracy in modeling the inverted pendulum, the dynamics and kinematics cannot be intimately correlated with the specific bipedal robot model. As a response, this paper introduces the concept of moment of inertia into the robot control process, significantly enhancing the robot's structural integrity. In contrast to the traditional centroid linear velocity control model, this approach incorporates moment of inertia, thereby improving feedback control effectiveness. The experimental results demonstrate that incorporating moment of inertia in this model enables stable responses during horizontal plane locomotion disturbances. Overall, this study introduces an innovative perspective for controlling retractable biped robot gaits.

Keywords: Robotics in intelligent manufacturing, Autonomous mobile robots and manipulators, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Grinding the pipe external surface using cup-shaped wire brushes is a widely adopted approach in pipeline maintenance. To achieve automated grinding of the external pipe, it is crucial to establish a grinding force model and explore the grinding mechanism accordingly. In this study, based on the characteristics of cup-shaped wire brushes, a controllable and efficient approach for grinding the pipe external surface is proposed. An experimental platform is constructed to simulate the grinding process and measure real-time pipeline grinding forces using a six-axis force sensor. The influence of different grinding parameters on the grinding forces on the curved surface of the pipe is investigated via parameter study. Furthermore, based on a single abrasive grain grinding force model, a predictive model for investigating grinding forces on the pipe external surface using is developed, where the fitted grinding force model obtained from the experimental data is taken into account.

Keywords: Artificial intelligence in robotics, Robotic vision and image processing, New theory and technology in robotics and biomimetics
Abstract: In recent years, the global robot market has witnessed substantial growth, particularly in the domain of service robots. Despite their expanding presence, service robots encounter limitations when operating autonomously in unstructured environments, primarily due to their constrained computational capacities. As a solution, the fusion of cloud and edge computing resources becomes imperative to expedite task inference and enhance scenario perception capabilities. The integration of cloud-edge-device models holds significant promise in bolstering the operational efficiency of robots. This entails the dynamic partitioning of intricate robotic tasks, executed collaboratively across cloud, edge, and device resources. In this landscape, deep neural network (DNN) models play a pivotal role in facilitating a wide array of robotic tasks. The inference time for each layer of a DNN model in actual deployment, emerges as a critical determinant in the model’s partitioning strategy. It also serves as an important metric influencing the model’s suitability for a specific hardware platform. This article presents an overview of recent advancements in predicting inference and training time of DNN models, summarizes the related methods, and finally discusses the challenges in this field and the research that can be studied in the future

Keywords: Artificial intelligence in robotics, Robotics in intelligent manufacturing
Abstract: Acquiring human skills offers an efficient approach to tackle complex task planning challenges. When performing a learned skill model for a continuous contact task, such as robot polishing in an uncertain environment, the robot needs to be able to adaptively modify the skill model to suit the environment and perform the desired task. The environmental perturbation of the polishing task is mainly reflected in the variation of contact force. Therefore, adjusting the task skill model by providing feedback on the contact force deviation is an effective way to meet the task requirements. In this study, a phase-modulated diagonal recurrent neural network (PMDRNN) is proposed for force feedback model learning in the robotic polishing task. The contact between the tool and the workpiece in the polishing task can be considered a dynamic system. In comparison to the existing feedforward neural network phase-modulated neural network (PMNN), PMDRNN combines the diagonal recurrent network structure with the phase-modulated neural network layer to improve the learning performance of the feedback model for dynamic systems. Specifically, data from real-world robot polishing experiments are used to learn the feedback model. PMDRNN demonstrates a significant reduction in the training error of the feedback model when compared to PMNN. Building upon this, the combination of PMDRNN and dynamic movement primitives (DMPs) can be used for real-time adjustment of skills for polishing tasks and effectively improve the robustness of the task skill model. Finally, real-world robotic polishing experiments are conducted to demonstrate the effectiveness of the approach.

Keywords: Applications of robotics and biomimetics, Robotics in intelligent manufacturing
Abstract: With the increasing use of special-purpose robots with heavy load requirements in narrow and critical environments , the operation safety and collision avoidance of such robots have become closely related, making trajectory planning an important research direction today. This paper proposes a path points interpolation method suitable for generating multi-objective optimized motion trajectories for specific tasks, thereby improving the overall performance of the robot, especially its operational manipulability. It mainly focuses on the joint space of the robot, optimizing its velocity, acceleration, position, and force. We use two evaluation functions to compare the performance of the modified B-spline method, and employ a genetic algorithm to find the best solution. Simulation and experimental demonstrations are conducted to validate the effectiveness of the trajectory optimization. The trajectory interpolation method we proposed can optimize various performance aspects of the robot while also providing a wider range of optimization options.

Keywords: Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics, Robotics in intelligent manufacturing
Abstract: The accuracy of the dynamic model can be seriously affected by dynamic noise, complexity of dynamic equations and coupling of parameters. To reduce this influence and the complexity of solving time and space, the gravity model of the manipulator is adopted. To obtain accurate identification of the manipulators gravity model, an identification method is proposed. Based on the least square method, the error cost function is designed to iterate the identification e rror to the minimum value (ITLSM). Given the process of gravity identification, a d ifference m ethod i s u sed t o e stimate the mass of the end effector and the position of the center of mass. Besides, an optimal pose based on the multi-objective DA algorithm is designed to optimize the estimation result of payload parameters. As demonstrated by experimental results, the maximum absolute identification e rror o f t he r obot joint gravity model is only 0.3 Nm, and the minimum error is 0.02 Nm. Moreover, the maximum absolute value of payload mass estimation error is 0.075 kg and the maximum absolute estimation error of payload centroid position is 0.052 m, verifying the effectiveness of the proposed method.

Keywords: Autonomous mobile robots and manipulators
Abstract: In this paper, we propose an efficient trajectory planning algorithm with path smoothing based on the Bézier curve with curvature constraints and piecewise-jerk speed-time optimization. We use hybrid A* to generate a rough path and construct a safe corridor by inflating the path. After that, we formulate the smooth problem as a nonlinear programming(NLP) with piecewise Bézier curves. Since the curvature constraints for Bézier curves are difficult, we employ quartic Bézier Curves with special forms and compute the closed-form solution for the maximum curvature to simplify the representation of the maximum curvature. By using the special Bézier curves, we realize the gear shifts and easily guarantee the security, continuity, and feasibility of the path. Meanwhile, we add time variables based on PJSO, improving the quality of trajectory within an acceptable increase in time, making the allocation of time and speed better. Simulation and real-world experiments with a car-like robot in various environments confirm that our algorithm can generate a smooth, feasible, and high-quality trajectory for robots.

Keywords: Applications of robotics and biomimetics, Autonomous mobile robots and manipulators
Abstract: In recent years, old bridges have increased and require inspection to detect any damaged parts and maintenance to decrease the risk of falling. Bridge inspection robots have developed rapidly to replace human labor. This paper presents a new development of an inspection robot using switchable permanent magnetic wheels rather than typical permanent magnetic wheels to control the adsorption force of each wheel, which can increase the maximum adsorption force of magnetic wheels compared to the previous design that needed to limit the adsorption force for successfully moving between surfaces. Furthermore, this robot has a total of six wheels that connect with leg structures to change its posture for moving to other surfaces on different paths. The movement process of the robot for changing the surface in each path is presented in this paper. As a result of the robot's performance, it can move and change to another surface in various paths and shows the potential to inspect the complex environment of the steel bridge.

Keywords: New theory and technology in robotics and biomimetics, Soft robotics and liquid-metal robotics, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: The body structure and hunting behavior of snakes have evolved through lengthy natural selection, making them highly remarkable. There are striking similarities between the soft robotics and biological snakes, including softness, adaptability, flexibility, and efficiency. Therefore, snakes are excellent bio-inspiration animals in the field of soft robotics. We propose a new mechanism called the Twisted String and Spiral Hose Mechanism (TSSH) that can be applied in the field of soft robotics, inspired by the snake's prey constriction. This mechanism can achieve complex locomotion with a simple component. It can deform into a helical shape, and the radius of the helix can be controlled. This mechanism has the potential to be applied in a griper that can disregard size and shape within a certain range. We conducted experiments to validate its performance. Subsequently, we employed a modeling method based on a discrete model for deformable linear objects to model and analyze the TSSH mechanism. At last, we proposed some potential future applications of this novel mechanism.

Keywords: New theory and technology in robotics and biomimetics, Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics
Abstract: The collaborative mode based on caudal and pectoral fins was utilized to achieve the standing-and-hovering (SAH) behavior in dolphins. A virtual 3D model consisting of the body, caudal fin and symmetrical pectoral fin is used to carry out numerical simulation studies instead of real dolphins. The hovering mechanism and the laws of variation in hydrodynamic performance parameters with time during the standing-and-hovering behavior were revealed. The results show that the standing-and-hovering behavior of dolphins could be achieved by the collaborative effect of the caudal and pectoral fins. The caudal fin and pectoral fins generate thrust alternately at 0.25, 0.75, and 0.5 cycles, respectively to maintain the stable standing-and-hovering behavior. For different desired objectives, suitable combinations of kinematic parameters can be chosen to achieve optimal results. This study provides a new approach to break through the spatial barriers to movement of underwater robots and provides a solid hydrodynamic theoretical basis for the development of cross-medium bionic robots.

Keywords: Applications of robotics and biomimetics, Autonomous mobile robots and manipulators, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Routine inspection and maintenance are crucial for safe and sustainable operation of wind turbine blades. Compared to repairing with human beings, climbing robots are safer and more efficient. There are some key issues in using robots to inspect and repair blades, such as crawling on surface with significant curvature changes, omnidirectional movement, etc. These issues demand high adaptability and movement capabilities of robots. Therefore, this paper presents a novel hexapod robot characterised with high adaptability and omnidirectional movement ability on blade surface. We designan underactuated leg structure and a dedicated controller and integrate it on proposed hexapod. The leg design enables robot adapt and attach to blade surface, and the controller enables robot crawl safely and omnidirectionally. Physical experiments demonstrate that the robot can adapt and move omnidirectionally on a blade surface.

Keywords: Applications of robotics and biomimetics, Space robots, aerial robots, and underwater robots, New theory and technology in robotics and biomimetics
Abstract: This paper investigates the key factors affecting the hydrodynamic performance of a class of planar linkage mechanism cycloidal propellers for self-propelled underwater robots using numerical simulation methods. Firstly, the kinematic models of the four-bar and mixed four-/five-bar mechanism cycloidal propellers are established, along with the physical model of the underwater robot equipped with the cycloidal propeller. Next, the effects of variations in eccentricity, control mechanism type, and revolution speed on the hydrodynamic performance of the cycloidal propeller are studied using Computational Fluid Dynamics (CFD) method. Finally, the influence mechanisms of these three factors on the hydrodynamic performance of the cycloidal propeller are discussed using wing element theory. This research explores the key factors for improving the hydrodynamic performance of self-propelled underwater robots equipped with cycloidal propellers, the results indicate that the changes in the three factors studied have a significant impact on the performance of cycloidal thrusters, providing important references for the design and optimization of cycloidal propellers for underwater robotics.

Keywords: Artificial intelligence in robotics, Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics
Abstract: 以降低地图的维护成本并增强 三维自主探索中的搜索效率 环境，本文提出了一种针对 空中机器人平台的大规模3D测绘。这 拟议规划师采用多地图协作方案， 其中对 范围有限、成本低的高精度子地图 全局地图提供子地图的信息来指导 粗略的探索方向。结果，边疆 增量搜索，而不是重复搜索 确定整个已知地图和最佳边界。在3D环境中，其性能由以下公式评估 与波前前沿（WFD）算法和Next算法的比较 最佳视图规划师 （NBVP） 算法。所提出的方法 减少 3% 和 20% 以上的探索时间 分别。

Keywords: Artificial intelligence in robotics, Autonomous mobile robots and manipulators, Multi-robot systems, swarm robots, and collaborative robots
Abstract: Multi-agent coverage path planning (MACPP) with unmanned aerial vehicles (UAVs) is an essential task in several real-world applications, such as environmental monitoring, surveillance, and search and rescue missions. Existing MACPP algorithms suffer from limitations, such as the inability to generalize over all system parameters, long computational time, and limited applicability in dynamic environments. In this study, we propose a novel MACPP framework based on proximal policy optimization (PPO), a model-free reinforcement learning algorithm. Our framework provides a dynamic 3D machine learning-based approach that can generalize over all system parameters, such as target zones, agent positions, battery levels, and other critical factors. Hence, the framework offers several advantages, including efficient and effective path planning, real-time application on hardware, and applicability in complex 3D structures such as buildings or alpine terrain. We demonstrate the feasibility and effectiveness of our approach by successfully applying it in real-time on hardware. Our experimental results show that our proposed framework can effectively handle 3D structures and react to changes mid-flight.

Keywords: Robotic vision and image processing, Artificial intelligence in robotics
Abstract: Insulator defect detection is important for the safety operation of the power grid, which can be inspected via the unmanned aerial vehicles (UAVs) patrolling, demanding high accuracy and real-time capability. In response to this requirement, this paper investigates a real-time end-to-end insulator defect detection algorithm, RT-DETR (Real-Time DEtection TRansformer) with the combination of the model compression method based on the parameter quantization and knowledge distillation. In order to reduce the model parameters and accelerate the detection speed, a lighter backbone and a regularization method for refining the attention computation are applied in the model. Further, a quantization training approach which combines the parameter quantization and self-distillation is used for the model compression. The proposed method is trained and validated on an open-source dataset. Experimental results demonstrate that the average mean average precision (mAP) of the proposed method for the insulator defect detection is 99.5%, and the inference speed is 23ms, meeting the requirements for the UAVs real-time inspection.

Keywords: Applications of robotics and biomimetics, Smart sensors and actuators
Abstract: In this paper, we adopted a VTOL UAV model design and developed a digital twin for the corresponding model. The digital twin consists of the following parts: Unity simulation system, physical VTOL UAV, and communication interface. As we know, the main function of the digital twin is verification. However, there is a gap between the virtual simulation and physical test results. Therefore, based on the Unity engine and the VTOL UAV, this paper will demonstrate a simulation case of performing short takeoff and vertical landing operations using the VTOL UAV. These simulation results show that the system can simulate the aerodynamics, physics, and controls of the UAV with high fidelity and reliability. The digital twin system can also provide a platform for VTOL UAV application development, theoretical investigation, and intelligent training.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Artificial intelligence in robotics, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Transportation of payloads in cooperative manner is common in human life. It involves the policy to suppress payload swing during legged walking, which is quite different from the wheeled transportation. To promote the field of swinging dynamics in multi-robotics system, we discussed the characteristic of two quadruped robots to deal with a tether-suspended payload based on proximal policy optimization reinforcement learning algorithms. Instead of relying on complex model-based control frameworks, the deployment of our method was relatively simple for dynamic walking. The control problem was modeled in simulation and a universal quadruped robot motion control reward mechanism was developed. Domain Randomization was adopt to minimize the Sim2Real gap. Two locomotion methods including omnidirectional walking and directional walking were demonstrated to tracking a rectangle path and an eight pattern path. And the policy was deployed on Go1 quadruped robot in real experiments.

Keywords: New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: The performance of dynamic control is intimately tied to modeling accuracy. However, traditional estimation methods and friction models, such as the least squares method and the Coulomb plus viscous model, fail to reflect the actual characteristics accurately. Particularly, the linear nature of the Coulomb plus viscous model overlooks the nonlinear static features of joint friction at slower velocities. To improve the rationality of the model structure, we integrate the Stribeck friction model into the Coulomb plus viscous model. However, introducing such nonlinearities compromises the applicability of the least squares method. As a countermeasure, we propose a new strategy that combines the least square method and the Quasi-Newton iterative method to identify the parameters of the modified nonlinear model. Additionally, the design of the excitation trajectory is critical to achieve high identification accuracy. We utilized the inverse of the smallest singular value of the observation matrix as the objective function. By minimizing it with the interior point method, we generate the excitation trajectory well-suited to stimulate dynamic characteristics. Then we leverage the discrepancies between the measured and estimated torques to assess the precision of the dynamic parameters of the manipulator. Remarkably, our proposed algorithm reduces the mean absolute error of the estimated torque by over 20.40%. Finally, an experiment of the industrial manipulator by hand guiding grab and drag is performed and shows that the proposed approach can provide the manipulator with comprehensive torque compensation.

Keywords: Human-robot interaction, Multi-robot systems, swarm robots, and collaborative robots, Artificial intelligence in robotics
Abstract: In modern scenarios, kinesthetic teaching helps users to quickly program and plan trajectories in complex environments without specialized knowledge of robots. It has become one of the typical applications of human-robot interaction (HRI) to improve efficiency. Since the dynamic model of the robot reflects the mapping between joint motion and driving torque, accurate models are required to achieve good performance in sensorless kinesthetic operation control. Unfortunately, due to model uncertainty, assembly errors, and a lack of information provided by manufacturers, it is difficult to obtain a reliable dynamic model. This paper proposes a novel robot sensorless kinesthetic teaching scheme based on sparse feature learning dynamics. Firstly, without physical structural parameters, precise robot dynamics are learned through data-driven technology. Secondly, a learning-model-based reference controller is designed to effectively reflect the motion state of the robot under external force. The sequential minimal optimization (SMO) algorithm is used to eliminate the adverse effects of static friction on traction. Finally, the control scheme is used to achieve sensorless kinesthetic teaching on a 7-degree-of-freedom robotic arm.

Keywords: Space robots, aerial robots, and underwater robots
Abstract: This paper addresses the time-optimal trajectory planning problem of the free-floating space robot. Due to the path dependent dynamic singularities, the direct kinematics equations are employed. The joint range, velocity and acceleration limits are considered. After transforming the trajectory planning problem into an unconstrained nonlinear programming problem, particle swarm optimization algorithm is used to find the optimal trajectory. In addition to the requirement for the terminal pose of the end-effector, the spacecraft attitude can also be constrained in different forms through non-holonomic constraints. Simulation results are presented for trajectory planning of a 6 degree-of-freedom (DOF) space robot and demonstrate the effectiveness of the proposed method.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Robotics in intelligent manufacturing, Human-robot interaction
Abstract: Accurate contact force estimation is essential for achieving safe and stable physical interaction between a robot and its environment. Currently, contact force estimation methods based on joint force signals present challenges such as poor accuracy and significant noise. This paper proposes the disturbance Kalman filter to improve the performance of contact force estimation. This algorithm estimates contact force using a joint force sensor based on a generalized momentum-based robot dynamic model. By employing polynomial functions to model disturbance dynamics and virtual joint forces in the system state, we achieve accurate contact force estimation through a Kalman filter. Furthermore, a parameter tuning method for the disturbance Kalman filter is proposed to enhance the robustness of the algorithm to process noise and model uncertainty. Finally, this method is compared with the Generalized Momentum Observer applied to 5-DoF collaborative manipulators. Simulation and experimental results demonstrate that the disturbance Kalman filter can provide a robust and accurate estimation of contact force.

Keywords: Autonomous mobile robots and manipulators, Applications of robotics and biomimetics
Abstract: Legged robots have the advantages of strong flexibility, good stability, and a strong ability to overcome obstacles. Due to the occlusion of the camera’s perspective, the legged robot cannot quickly avoid obstacles when planning a path. This paper proposes a visual navigation framework for legged robots to plan trajectories and accurately follow trajec tories in unknown environments rapidly. The entire navigation framework includes two aspects. Regarding trajectory planning, we introduce dynamically adjusted weight coefficients and modify the search direction of the A* algorithm. Regarding the trajectory following, we dynamically calculate the look-ahead distance based on the current speed and yaw value to better fit the trajectory. We integrate the entire visual navigation framework into a legged robotic system. Experimental results show that applying dynamic look-ahead distance enables the robot to fit the trajectory better, improving the accuracy and stability of navigation. The improved A* algorithm is superior to the original method in the consumption of search time and the number of traversed nodes, which further improves the planning efficiency. In the simulation experiment, the search time is reduced by 38%, and the number of search nodes is reduced by 57%. In real-world experiments, the search time is reduced by 17%, and the number of search nodes is reduced by 30%. The framework proposed in this paper provides ideas for the rapid search path of legged robots in complex environments. It is of great significance to the practical application of legged robots.

Keywords: Autonomous mobile robots and manipulators, Artificial intelligence in robotics, New theory and technology in robotics and biomimetics
Abstract: We propose a new nonlinear optimization-based method for signal temporal logic (STL) specification with uncertain external events. STL provides a simple way to express complex task specifications for robotic and cyber-physical systems. Event-based STL is an extension of STL and can describe the response to external events, which should be considered in practical applications. We define smooth robustness functions for event-based STL specifications and formulate a nonlinear optimization problem for event-based STL specifications. Our method introduces augmented continuous variables for external events into the optimization problem to generate robust trajectories for uncertain external events that may occur. We demonstrate the effectiveness of our method through numerical simulations.

Keywords: Autonomous mobile robots and manipulators, Artificial intelligence in robotics, New theory and technology in robotics and biomimetics
Abstract: Trajectory planning involves generating a series of space points to be followed in the near future. However, due to the complex and uncertain nature of the driving environment, it is impractical for autonomous vehicles (AVs) to exhaustive design planning rules for optimizing future trajectories. To address this issue, we propose a local map representation method called Velocity Field. This approach provides heading and velocity priors for trajectory planning tasks, simplifying the planning process in complex urban driving scenarios. The heading and velocity priors can be learned from demonstrations of human drivers using our proposed loss functions. Additionally, we developed an iterative sampling-based planner to train and compare the differences between local map representation methods. We investigated local map representation forms for planning performance on a real-world dataset. Compared to learned rasterized cost maps, our method demonstrated greater reliability and computational efficiency.

Keywords: Artificial intelligence in robotics, Robotics in intelligent manufacturing
Abstract: The manipulation of deformable linear objects (DLOs) by robots is challenging because of the complexity of modeling DLO dynamics. Although previous studies generally employed physical models and data-driven approaches to predict DLO deformations, only a few studies have considered the contact of DLOs with the environment. In this study, we propose a framework integrating differentiable simulations with neural networks (NNs) to generate manipulation trajectories that avoid contact and achieve the goal shape. First, we implement a differentiable simulation to simulate the deformation and interaction of DLOs via position-based dynamics. Thereafter, we utilize the backpropagation of losses from the differentiable simulation to optimize the parameters affecting the deformation of DLOs in the simulator and explore an ideal manipulation trajectory for the task via an NN controller. The simulation and real-world experimental results reveal that the proposed method can generate valid manipulation trajectories from offline learning, which can also function well in real world applications using the optimized parameters.

Keywords: Artificial intelligence in robotics, Applications of robotics and biomimetics
Abstract: Multi-gait locomotion is an essential component of various quadruped robot control tasks. However, learning multi-gait locomotion control is challenging due to the considerable differences among gait motions. Meanwhile, these differences make motion imitation difficult to apply to learning multi-gait locomotion control. This work proposes a novel Decomposed Multi-Task Reinforcement Learning (DeMTRL) algorithm that alleviates the difficulty of imitating multiple gaits. The DeMTRL algorithm imitates the decomposed multi-gait leg reference motions with decentralized reinforcement learning. The decentralized learning scheme will exploit the similar leg motion pattern and reduce the difference in imitating different gaits. Additionally, the DeMTRL algorithm can enhance gait stability by considering the morphological symmetry of quadruped robots. Simulation results show that the DeMTRL algorithm can obtain stable real-time multi-gait locomotion and transition.

Keywords: Robotics in intelligent manufacturing, New theory and technology in robotics and biomimetics
Abstract: In order to improve the poor accuracy of industrial robots, the calibration technique for robot geometric parameters has emerged as a mainstream approach. However, existing calibration algorithms do not perform as expected due to initial large residual errors.To solve the problem, this work proposes a hybrid LR-QNLM algorithm based on the modified quasi-Newton (MQN) and Levenberg-Marquardt-Fletcher (LMF) methods.Firstly, we build the robot kinematic error identification model considering extended axes as the objective equation.Then, hybrid strategies are used to incorporate two methods above for a more accurate approximation of the Hessian matrix and well-local convergence property.Moreover, a non-monotonic linear search algorithm is used to achieve sufficient descent ability with a control criterion used to achieve more precise iteration.Finally, calibration experiments show that the robot positioning accuracy and robustness are well improved, which verifies the feasibility of the proposed method.

Keywords: Applications of robotics and biomimetics, MEMS, NEMS, nano-technology, and micro/nano systems, New theory and technology in robotics and biomimetics
Abstract: Benefiting from the decent converse piezoelectric effect, the piezoelectric nanopositioning stage has been applied to various micro-to-nano motion scenarios with splendid control performance and advantages. However, the inherent nonlinearity characteristics of piezoelectric significantly degrade the tracking accuracy and result in obvious positioning uncertainties. Especially the hysteresis effect mainly brings about the apparent control period lag and the reciprocating stroke inconsistency. Substantial existing control methods focus on establishing approximate hysteresis or inverse hysteresis model to suppress these adverse nonlinearities. Nevertheless, these methods commonly require sophisticated nonlinear modeling procedures, and the control performance largely relies on the modeling accuracy. Notably, the linear characteristics of the piezoelectric actually still perform a major influence on the control process, which can be significantly suppressed by straightforward linear algorithms. After compensating for the linearities, the nonlinearities only demonstrate an obviously marginal impact on the control performance. On this basis, the linear and nonlinear methods integration idea is proposed to replace the conventional modeling algorithm in this paper. Several combinations are discussed and conducted on the stage to validate the effectiveness of this idea.

Keywords: Autonomous mobile robots and manipulators, Artificial intelligence in robotics
Abstract: Hyper-redundant manipulators have superior dexterity. However, conventional control based on accurate modeling is difficult to achieve. This paper proposes a control method for the hyper-redundant manipulator to complete target approaching tasks using deep reinforcement learning. A hyper-redundant manipulator simulation model was first built and the corresponding forward kinematics equation was analyzed. Then, we used reinforcement learning methods to construct controllers for single target tasks, during which the impact of several key factors on the training was analyzed and optimized accordingly. On this basis, the multi-target task training was achieved by expanding the state to a generalized state. Finally, the proposed control method was verified in a dynamic simulation environment. Simulation results show the effectiveness of the proposed control method.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Artificial intelligence in robotics, New theory and technology in robotics and biomimetics
Abstract: The dynamics of flapping wing flight is notoriously challenging to model due to its inherent complexity. This paper presents the dynamics learning method based on Gaussian Pro- cess Regression (GPR), and the further implementation in tra- jectory tracking controller development. Specifically, through actively querying selected instances for labels, the algorithm can learn more efficiently with fewer labeled examples. Meanwhile, in order to improve the efficiency of the active learning process, a batch-wise data selection approach is employed. By taking advantage of the learned Gaussian Process Regression model, an optimization-based control strategy is proposed to refine actuator inputs using acquired knowledge. The superiority of the proposed control strategy are validated by high-fidelity numerical simulations

Keywords: Human-robot interaction, Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics
Abstract: Predictive motion planning enables robots to consider workspace obstacles and proactively avoid them. Estimating future poses of various obstacles is complex and exhibits results with uncertainties. Especially in the case of human-robot collaboration, the forecast process has to deal with sudden motion changes and varying human intentions. This paper proposes an uncertainty estimation for human motion extrapolations in the joint space of a human skeleton. The approach utilizes Gaussian Mixture Models that allow an online determination of the accuracy of current extrapolations based on preceding extrapolation errors. Multiple human model representations enable consideration of estimated uncertainties and are presented along with evaluating the resulting robot trajectories. The results show that considering estimated uncertainties leads to a robot trajectory with increased distances between humans and robots and decreased danger index from the literature, which is desirable for collaborative scenarios.

Keywords: Human-robot interaction, Robotics in intelligent manufacturing, Applications of robotics and biomimetics
Abstract: Robotic constraints avoidance is important for collaborative robots to achieve safe interactions with operators. However, it is difficult to simultaneously handle the various constraints which are described in different space. In this research, a control framework is proposed for solving the problem of robotic constraints avoidance based on virtual fixtures and adaptive admittance control. Specifically, different constraints are evaluated by geometric distances and redescribed as potential fields in Cartesian space. Then, a virtual repulsive force is generated with considering operator's intention to drive robot away from the constraint areas. Furtherly, an adaptive admittance controller is constructed to ensure the stability of system. Simulations are conducted on AUBO I5 robots under three scenarios and the effectiveness of the proposed framework in handling robot's self-collision, workspace collision, configuration singularity and joint-limit violation is verified.

Keywords: Autonomous mobile robots and manipulators, Artificial intelligence in robotics, Robotics in intelligent manufacturing
Abstract: In the ALFRED challenge for robot simulation, the robot still faces a challenge to schedule a task in the embodied instruction following (EIF) tasks. These tasks require the robot to accurately perceive visual features and understand language instructions. However, the previous approaches typically employed end-to-end structures that utilize a shallow understanding of language instructions, while EIF tasks demand a deeper understanding of the semantic relationships in these instructions. To overcome these limitations, we propose a method which named REIF. The method incorporates modules for visual perception, language understanding, semantic search, closed container prediction, navigation, and operation to form a modular framework based on visual language multi-modal learning. The semantic search module supports more efficient object search, while the closed container prediction module enables deeper language understanding. Through learning multiple task instructions, the robot can efficiently and accurately complete EIF tasks in unseen scenes under certain step lengths. Our framework performs significantly well on unseen scene tasks within the ALFRED benchmark, achieving state-of-theart accuracy and efficiency rates of 50.83% and 23.06% respectively. These results demonstrate that our method is capable of efficiently and accurately inferring the presence of closed container in unseen scenes, and can successfully execute a series of actions to interact with target object within closed container. Our method has achieved the first place in the ALFRED data-set competition. You can find our submissions and results at the following link: https://leaderboard. allenai.org/alfred/submissions/public.

Keywords: Autonomous mobile robots and manipulators, Applications of robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Wheel-legged robots are widely used because they combine the ability of wheeled robots with high locomotion efficiency and legged robots with high environmental adaptability. However, it is important and challenging for wheellegged robots to rationally plan the switching between wheeled and legged locomotion modes. In this paper, we propose a planning model based on energy optimization to solve the locomotion mode switching process with the self-developed wheel-legged robot (Mobot-WL) as the research object. The core idea is to solve a smooth and less energy-consuming locomotion mode switching trajectory based on the stabilization of the motion process. Simulation results and experiments with Mobot-WL show that the algorithm can plan the locomotion mode switching efficiently and robustly.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Artificial intelligence in robotics
Abstract: Quadruped Robots require the capability to traverse complex and demanding surroundings found in natural landscapes, urban areas, and industrial facilities. However, achieving efficient and flexible control of these systems remains an ongoing challenge. In this paper, we present a novel Universal Dynamic Model Predictive Control (UDMPC) Framework designed specifically for quadruped robot locomotion, aiming to address the controller design problems caused by the diversity of quadruped robots and the complexity of terrains.Even within intricate terrains, this framework guarantees accurate tracing of reference velocity commands, augmenting locomotion prowess. The MPC controller's parameters are dynamically fine-tuned through reinforcement learning, enhancing control reliability. Our proposed approach was subjected to rigorous testing and evaluation using the Go1 quadruped robot model within a simulation environment. The findings showcased its exceptional dynamic adaptability, surpassing fixed-parameter controllers. Notably, this work considerably enhances command tracking precision and stability.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics
Abstract: When the quadruped robot is engaged in logistics transportation tasks, it encounters a challenge where the distribution of the center of mass (CoM) of the loaded items is not only random but also subject to time variations. Consequently, the robot becomes susceptible to non-zero resultant torques, which inevitably impact its body posture during the walking process. This paper proposes a method to estimate the CoM inertia using four one-dimensional force sensors and a walking control strategy for complex urban terrain. The inertia tensor and CoM of the load are first estimated, then the robot’s dynamics are compensated, and foothold adjustments are made for underactuated orientations to compensate for the extra moment generated by the CoM offset. For uneven terrain, the terrain estimator and event-based gait are used to adjust the robot's gait to reduce the impact of terrain changes on the robot. The effectiveness of the proposed method and the feasibility of load walking in urban terrain are verified through comparative experiments, complex terrain load walking experiments in Webots, and real prototype experiments.

Keywords: Autonomous mobile robots and manipulators, Robotic vision and image processing, Artificial intelligence in robotics
Abstract: The occurrence of catastrophic caused tons of thousands of lives. To enhance the efficacy of search and rescue operations in challenging to access region, we have introduced an autonomous hexapod robot equipped with a partial human detection and tracking system that is capable of identifying fragments of the human body. The chosen hexagonal arrangement of six legs enables the robot to navigate through rough terrains. This configuration provides stable mobility, with an average acceleration of 0.8 m/s² and a peak acceleration of 3.1 m/s² during walking, making the camera of the robot remains steady for human identification. Due to the employed DeeplabV3+ architecture for the training process, partial human bodily structure detection has achieved 94% accuracy. Moreover, the robot can track the detected individual by maintaining 1 m distance.

Keywords: Multi-sensor data fusion and sensor networks, Human-robot interaction, Applications of robotics and biomimetics
Abstract: Person-following technology is an essential part of the robot's human-machine interaction function and has a wide range of application scenarios. However, tracking a single person in the real world through sensors still faces many challenges, e.g., in dense crowds, the tracking target can be partially occluded or temporarily missed. In this paper, we proposed a robust person-following framework based on LiDAR-camera fusion to solve these problems. The framework consists of three main parts. Firstly, the person detection results are obtained from both sensors by using different detection methods. Secondly, a cascaded spatial-temporal data association method is used to filter out the currently tracked person. Finally, the target observations from both sensors are fed into different Kalman filters to update the target motion state in different observation coordinate systems. Experiments on KITTI and our tracking dataset show that, even under partial occlusion, our method can still track the target continuously. Moreover, the experimental results demonstrate that our method can be implemented in real-time with an average processing frame rate of 15 Hz.

Keywords: Artificial intelligence in robotics, Human-robot interaction, Applications of robotics and biomimetics
Abstract: With the popularization of educational robots in the field of subject education, primary and secondary school students’ expectation for educational robots has also risen from the ordinary robot doll companion learning to the educational application of humanoid robots, hoping to learn in class with humanoid robots. In order to meet students’ demand for humanoid robot teachers’ interaction ability and English teaching function, we developed a set of humanoid robot English teaching system for primary and middle school students. The system has three functional modules: automatic teaching, intelligent question-answering and oral practice. The automatic teaching module imitates the classroom teaching of human teachers, and intelligent question-answering can enable robot teachers to answer students’ questions based on texts, which can be applied to the humanoid robots independently developed in our laboratory. At the same time, since most of the public use subjective evaluation indicators for educational robots, we designed an evaluation data set for the system, and through the designed evaluation method, the rationality of the system’s automatic teaching function and the accuracy of the intelligent question-answering function were objectively verified.

Keywords: Robotic vision and image processing, Human-robot interaction, Artificial intelligence in robotics
Abstract: Pedestrian tracking is an important research direction in the field of mobile robotics. In order to complete tasks more efficiently and without hindering the original intention of pedestrians, mobile robots need to track pedestrians accurately in real time. In this paper, we propose a real-time RGB-D pedestrian tracking framework. First, we propose a pedestrian segmentation detection algorithm to detect pedestrians and obtain their two-dimensional positions. Second, due to limited computational resources and the rarity of missed detection for pedestrians, we use an nearest neighbor tracker for pedestrian tracking. To address the issue of inaccurate pedestrian localization, we use our detection algorithm to obtain the center of pedestrians from RGB images. By combining them with point clouds, the 2D coordinates of pedestrians are obtained. Our method enables accurate pedestrian tracking in the world coordinate, by adaptively fusing RGB images with their corresponding depth-based point clouds. Besides, our light-weight detection and tracking algorithm guarantee the real-time pedestrian tracking for realistic mobile robot applications. To validate the effectiveness and real-time performance of tracking algorithm, we conduct experiments using multiple pedestrian datasets of approximately half a minute in length, captured from two different perspectives. To validate the practicality and accuracy of the tracking algorithm in real-world scenarios, we extend our tracking algorithm to apply it to trajectory prediction.

Keywords: Human-robot interaction, Autonomous mobile robots and manipulators
Abstract: This study focuses on investigating an adaptive impedance control strategy for facilitating human-robot collaborative tasks. Traditional impedance control methods may lead to significant position tracking and force tracking errors when accurate environmental information is not available. A force feedback-based impedance control algorithm is proposed to enhance control accuracy. The algorithm utilizes a neural network adaptive control approach to adjust the parameters of the impedance control algorithm, as well as estimate and compensate for unknown environmental information. The proposed control approach utilizes radial basis function neural networks (RBFNN) to enhance tracking performance by compensating for dynamic uncertainties. The proposed approach was tested under various conditions to ensure its performance across different situations.

Keywords: Autonomous mobile robots and manipulators, Robotic vision and image processing, Artificial intelligence in robotics
Abstract: The evolution of robots is continuously advancing to greater heights. Virtual reality as a tool provides an immersive experience in a 3D environment that allows users to visualise and interact with robots in such environments. This human-robot interaction benefits users with better situational awareness to effectively perform remote tasks. This paper proposes a seamless teleoperation pipeline that interfaces virtual reality with robots for precision agriculture tasks. This pipeline allows users to visualise and control robots from a virtual environment thereby giving a telepresence advantage to users. Robotics Operating System (ROS) is utilised to accept data from a Unity gaming plat- form. A jackal robot and panda robotic manipulator equipped with the required sensors are utilised to evaluate this pipeline. The camera data, odometry data, velocity data and joystick commands are interchanged between the virtual environment and real platforms. Further evaluation is conducted for 6DoF pose application where a target fruit pose is estimated and manipulated within the virtual world while imitating similar actions in the real world to grasp/harvest the fruit. This attempt exhibit the possibility of remote agricultural task through this pipeline. The results of the evaluation prove the efficiency and robustness of the pipeline.

Keywords: Human-robot interaction, Medical robotics, biomedical and rehabilitation engineering, Applications of robotics and biomimetics
Abstract: Machine-human interaction systems have been proposed to improve motion learning efficiency by providing feedback on motion misalignment between a learner and an instructor. Conventional motion teaching systems based on haptic information presentation generally use electrical sensors and motors, which causes the exoskeleton suit weight and the scale of the entire system to become large. In this study, we proposed a pneumatic-driven motion teaching system that provides feedback to the learner by simultaneously presenting visual and torque information to the learner. We achieved a lightweight, soft, and user-safety haptic system using pneumatic artificial muscles (PAMs) as actuators. PAMs' shrink force generates external torque on the learner's joint to correct the elbow flexion and extension motion misalignment between the instructor and the learner. We conducted a motion teaching experiment to verify the effectiveness of the proposed method. Specifically, we performed motion instruction on eight subjects using three patterns: a visual-only presentation method, a conventional method that simultaneously presents visual and vibrotactile presentation, and the proposed method that simultaneously presents visual and torque. The experimental results showed that the proposed method reduced the angle and angular velocity tracking errors compared to the visual-only method and visual-vibrotactile method.

Keywords: Human-robot interaction
Abstract: For the recent years, the authors have proposed an interface called Real-World Clicker (RWC) and have been developing a technology to intuitively indicate the position and posture of objects in everyday space to a support robot. Our RWC is a device in which a TOF-type laser distance sensor is mounted on a pan-tilt actuator and manually operated by a user with a PC mouse or a smartphone. In this paper, we reported pointing performance evaluation of smartphone-based operation methods. Specifically, four methods were compared; (A) PC mouse, (B) method that links the smartphone posture to the RWC posture, (C) 2-step tuning method with coarse and fine tuning, and (D) method that optimizes the fine tuning sensitivity based on the manipulability of a robot arm. Six pointing tasks were given to 5 subjects and their pointing performances were evaluated and analyzed using Throughput value. As a general result, we concluded that (D) was superior to (B) and (C) and comparable to (A).

Keywords: Medical robotics, biomedical and rehabilitation engineering, Applications of robotics and biomimetics
Abstract: Dependency on visual contribution to balance ability is a common problem for elderlies. Therefore, there is necessity to develop a training system that aims to reduce visual dependency and smoothly transfer to somatosensory sensory to maintain postural control. In this research, we developed a standing balance training system that consists of a Wii board to measure center of pressure (CoP) position while standing, a virtual reality (VR) headset that present immersive tilt image, and a vibrotactile biofeedback (BF) belt. An experiment was conducted on twelve young and healthy participants to verify feasibility of the proposed system. Six participants randomly assigned to BF group were provided with vibrotactile BF of correct CoP position for standing posture correction while seeing tilt image. The other six participants assigned to NBF (No biofeedback) group were only presented with tilt image. The within group results show that standing stability while seeing normal image (p = 0.028) and tilt image (p = 0.028) significantly improved for BF group, while those of NBF group had no significant differences (p = 0.345 and p = 0.075, respectively). For between group results, standing stability for BF group was significantly better than that of NBF group while presenting both normal image (p = 0.046) and tilt image (p = 0.028). Such results indicate the feasibility of the proposed system to reduce visual dependency. Especially, the vibrotactile BF made adjusting ratio of visual and somatosensory contribution to standing postural control smoother. A pilot study in elderly people is left as a future direction for the research.

Keywords: Human-robot interaction, Applications of robotics and biomimetics, Artificial intelligence in robotics
Abstract: This article proposes an approach to modeling a human driver using a recurrent neural network (RNN). The obtained model is used to reproduce the driver behaviors in a human-in-the-loop system, in which the human driver acts as a positioning controller by performing steering operation during driving. In the proposed method, the RNN is incorporated in the closed-loop system as a controller, and the controlled output of the closed-loop system operated by a human driver is fed into the RNN as the teaching signal for training. This approach ensures the stability of the closed-loop system, unlike when the RNN is first trained using the input and output signal of the RNN itself and later have the obtained model incorporated into the closed-loop system. Training data were acquired through two experiments where the driver was asked to perform positioning tasks with either no display of the target trajectory or with a two-second preview of the target trajectory. The latter experiment was performed so that the RNN can learn human predictive characteristics in addition to the actions as a feedback controller. Results show that the RNN was able to reproduce human actions.

Keywords: Robotic vision and image processing, Human-robot interaction
Abstract: As an artful movement, dance demands performers to execute precise actions and maintain graceful postures. However, evaluating a dancer's movements can be highly subjective and cumbersome. To address this issue, our study proposed an intelligent method for dance movement recognition and assessment based on human posture estimation. Regarding movement recognition, we refined the original COCO skeleton model, followed by a lightweight improvement to the OpenPose model. Compared to the original model, the parameter quantity decreased by 88.9%. We employed the improved OpenPose algorithm to extract key points of the human body from videos and used Singular Spectrum Analysis (SSA) to denoise the data. For the assessment aspect, we utilized the Dynamic Time Warping (DTW) algorithm to measure the similarity between actual dance movements and standard dance movements in terms of joint angles, thereby assessing the dance movements. This method can provide dancers with real-time feedback and guidance, objectively evaluating their dance movements. The proposed intelligent method can be used in controlling humanoid robotics or rehabilitation robotics due to the accurate human pose estimation and movement assessment results.

Keywords: Human-robot interaction, Medical robotics, biomedical and rehabilitation engineering
Abstract: Ideally, intelligent surgery should allow surgeons to achieve insight, prediction, and assessment of the full range and course of diseased tissue without causing perceptual and interactive inconvenience. However, in the clinical setting, surgeons can only achieve judgment of the condition by continuously exposing patient tissues and organs through dissection, which is risky. To address these issues, this paper proposes a surgical human-machine natural interaction assistance system based on Holographic Reality. First, we annotate and segment the patient's medical images and plan the surgical path to generate a knowledge-based 3D digital human model containing information about the tissue structure of the lesion and the access to the surgical instruments. Secondly, we introduce techniques such as motion recognition and force feedback into the virtual reality environment for holographic display and somatosensory interaction, achieving gesture interaction and virtual surgical instrument manipulation of the digital manikin. This system uses medical image processing and holographic reality technology to help surgeons perform preoperative surgical planning and intraoperative real-time information guidance to obtain accurate and safe surgical results.

Keywords: Human-robot interaction, Robotic vision and image processing, Artificial intelligence in robotics
Abstract: Human pose and shape estimation is crucial in perceiving gestures and actions intended by the human. It is necessary to perform this task in real-time so that correctness and safety can be carried on. It is also important that this task is executable at practically affordable machines. We thereby come up with the integrated pipeline for robustly perceiving human poses and shapes on an edge device. The pipeline is a combination of YOLOv8, AlphaPose, MotionBERT, and ExPose. Using these models, the pipeline performs bounding box detection, 2D skeleton estimation, 2D-to-3D pose lifting, and 3D mesh reconstruction, respectively. Results of the related experiments show that our pipeline can accurately and precisely process input images in real-time on the Jetson TX, a typical example of an edge device.

Keywords: Human-robot interaction, Multi-sensor data fusion and sensor networks, Artificial intelligence in robotics
Abstract: In this paper, we hypothesize that human relationships can be estimated from interactions and the frequency and length of mutual gazes between an individual pair in videos, particularly those captured by surveillance cameras at convenience stores, supermarkets, and shopping malls. Recently, there has been significant research progress in human interaction recognition for video surveillance systems. However, mutual gaze detection in surveillance camera’s views still remains a challenge. To verify our hypothesis, we collected a simple mutual gaze dataset at our laboratory and developed a system for mutual gaze detection built on top of a gaze estimator. We then collected a dataset with two types of relationships in the replica room of a convenience store at our laboratory, deployed our developed mutual gaze detector and existing human interaction recognizer, and introduced a human relationship recognition model with a self-attention module. The results suggest that the model pays attention to detected mutual gaze signals and predicted interactions as it learns to classify the human relationships.

Keywords: Applications of robotics and biomimetics, Human-robot interaction, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract:

The growing demand for electric vehicles requires the development of automated car charging methods. At the moment, the process of charging an electric car is completely manual, and that requires physical effort to accomplish the task, which is not suitable for people with disabilities. Typically, the effort in the automation of the charging task research is focused on detecting the position and orientation of the socket, which resulted in a relatively high accuracy, ±5 mm and ±10 degrees. However, this accuracy is not enough to complete the charging process. In this work, we focus on designing a novel methodology for robust robotic plug-in and plug-out based on human haptics to overcome the error in the orientation of the socket. Participants were invited to perform the charging task, and their cognitive capabilities were recognized by measuring the applied forces along with the movements of the charger. Eventually, an algorithm was developed based on the human's best strategies to be applied to a robotic arm.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Smart sensors and actuators, Human-robot interaction
Abstract: Computer vision-based methods for human pose estimation often encounter challenges such as occlusion and the impact of environmental lighting. The use of RFID technology and wearable flexible tags can overcome these challenges. Controlling the antenna's movement not only expands the detection range of the antenna but also applies to scenarios where the antenna is installed on a mobile care robot for human body pose estimation. In this paper, we propose a method to detect human posture by using the mobile antenna and RFID technology. Firstly, the collaborative robot controls the antenna movement. The coarse-grained position of the joint tags is quickly obtained using the non-linear least-squares method. Then, we iteratively improve the initial estimates based on the coarse-grained positions to derive the fine-grained positions of the tags to model the human skeleton. The limb rotation angle is estimated by extracting two phase features: the phase difference of different tags of the same antenna (PDT) and the phase difference of the same tag of different antennas on the limb (PDA). We implement and evaluate the RF-MvCare prototype using commercial RFID readers and tags, comparing its human pose estimation performance with Microsoft's Kinect 2.0, which serves as the benchmark. The experimental results demonstrate that RF-MvCare achieves an average limb angle error of 5.5° and an average joint position error of 3.6 cm for human pose estimation, indicating its accuracy and potential for applications in medical and home care settings where non-line-of-sight (NLOS) sensing is crucial.

Keywords: Human-robot interaction, Medical robotics, biomedical and rehabilitation engineering
Abstract: Human-machine interfaces (HMIs) require minimal size and less interference to the users. Wireless and battery-free design of HMI devices can be promoted by wireless power transmission (WPT) technology. However, researchers often ignore the curvature of WPT coils. In addition, power transfer efficiency (PTE) and robustness against coil shifts are not both considered when designing WPT links. In this paper, a coaxial cylindrical coils model was presented to simulate curved coils, and several design parameters affecting robustness were investigated through electromagnetic and circuit simulations. We also proposed a novel method for designing WPT links that ensures both sufficient robustness and optimal PTE. Three sets of coils designed with the proposed method achieved an efficiency over 85% in simulations, and fulfilled the given robustness requirements at the same time. The coaxial cylindrical coils model and the proposed method were validated by experiments. A demonstration of the wireless powering of an HMI device was performed as well. In conclusion, this paper provided a guideline to model and design efficient and robust WPT coils in HMI applicaitions.

Keywords: Human-robot interaction, Autonomous mobile robots and manipulators, Artificial intelligence in robotics
Abstract: The robot could provide convenient service for humans by following them. But it is not easy for the robot to follow a person while avoiding various objects in real crowded corridors. In this paper, a self-avoidance model predictive controller (SA-MPC) for human following in crowded environments is proposed. An obstacle avoidance optimization item is designed to enable the generalized controller to avoid collisions. Another adaptive method of selecting waypoints is introduced to make the robot move at the specific linear and angular velocities. We verified our method on the pedestrian simulation platform. Qualitative and quantitative simulations demonstrate that the proposed SA-MPC achieves a higher scene pass rate and improves safety and robustness in dense environments. Simulations are implemented in dense corridors and complex scenarios with random obstacles. SA-MPC can serve as a generalized controller for various mobile robots.