Keywords: Human-robot interaction, Applications of robotics and biomimetics
Abstract: Nonlinear stiffness SEAs (NSEAs) inspired by biological muscles offer promise in achieving adaptable stiffness for assistive robots. While assistive robots are often designed and compared based on torque capability and control bandwidth, NSEAs have not been systematically designed in the frequency domain due to their nonlinearity. The describing function, an analytical concept for nonlinear systems, offers a means to understand their behavior in the frequency domain. This paper introduces a frequency domain analysis of nonlinear series elastic actuators using the describing function method. This framework aims to equip researchers and engineers with tools for improved design and control in assistive robotics.

Keywords: Soft robotics and liquid-metal robotics, Smart sensors and actuators, Applications of robotics and biomimetics
Abstract: The compliance nature of soft robots permits them to effectively adapt to various challenging environments, but at the cost of poor payload carrying capabilities and lack of precision in motions. Electro-adhesive (EA) layer jamming clutches are an emerging type of active variable stiffness (VS) technique that features the advantages of structural simplicity, light-weight, rapid response speed, and low power consumption. However, current EA clutches mainly adopt a planar configuration, despite the large stiffness variations in tensional direction, they fail to exhibit effective stiffness variations in compression and bending directions due to the planar structures. To address this drawback, a novel corrugated EA clutch is proposed in this work to realize multi-directional VS. Its VS performances in tension, compression, and bending directions are benchmarked against a widely studied planar EA clutch and the effects of actuation voltage amplitude are investigated experimentally. The proposed corrugated EA clutch can effectively increase the brake forces when activated by 25, 23 and 8 folds in tension, compression and bending directions, respectively. To demonstrate its applications in soft robotics, a novel origami gripper with excellent adaptability is proposed by utilizing the corrugated EA clutch, and successful grasping on objects with various weights and morphologies is demonstrated.

Keywords: Robotics in intelligent manufacturing, New theory and technology in robotics and biomimetics, Autonomous mobile robots and manipulators
Abstract: Abstract—Lightweight design can improve the energy efficiency of biped robots, which is significant in the field of biped robot. Topology optimization has been widely used in the field of robot. However, the existing optimization is limited to the structural space based on traditional empirical design, and the potential of topology optimization is not fully utilized. Moreover, the finite element model is too simplified, which leads to a large optimization deviation. In this paper, a method of structural innovation design for the shank structure of biped robot is proposed. In the initial stage of conceptual design, the design space of optimization object is effectively maximized. Then, the load transmission characteristics of bearings are simulated by link10 element, and the topology optimization method of multiple load cases is carried out to obtain the optimized shank structure, which reduces the weight by 14.09 % compared with the existing shank. Finally, finite element extreme strength check and experiment test are both used to verify the feasibility and availability of the structure.

Keywords: Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: In contrast to head-first swimming of most natural swimmers such as fishes, the larvae of mosquito swim tail-first with the posterior end of the body leading the anterior end in the direction of locomotion. In order to gain comprehensive insights into the mechanics of tail-first swimming, the gait kinematics was quantified through markerless tracking, and flowfield was computed using Particle Image Velocimetry (PIV) analysis, revealing that the generation of peak hydrodynamic forces was at mid-stroke. Parameterization of the kinematics facilitated self-propelled numerical simulations using immersed boundary lattice Boltzmann method (IB-LBM) to be performed and the implementation of kinematics on a tethered robotic model. The parameterization produced tail-first swimming in the numerical simulations and the dye flow visualization from robotic experiments reveals the direction of momentum transfer facilitates tail-first locomotion. Furthermore, force measurements on the robotic model provide quantitative evidence supporting tail-first swimming resulting from this gait.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Applications of robotics and biomimetics
Abstract: The Oropharyngeal (OP) swab sampling is a routine, simple and effective method for the detection of respiratory infectious diseases, especially when a large-scale infectious disease outbreak occurs, such as the newly experienced COVID-19. The traditional manual OP swab sampling process is full of difficulties, such as the high demand for professional medical staff, the risk of cross-infection, and the inflexibility of sampling time, etc., making the efficient implementation of this sampling method particularly critical. Therefore, this paper proposes a novel OP swab sampling robot system to assist the OP swab sampling task. The dual-stage series-parallel hybrid manipulator component is driven by four flexible shafts with torsion properties, which guarantees the completion of sampling tasks in the narrow space of the oral cavity. Accurate sampling point location information is provided through sensor calibration and location recognition algorithms based on deep learning, and the entire sampling task is implemented using a visual-tactile fusion control system. The preliminary experiments verified the completion and feasibility of the sampling task of the OP swab sampling system, and provided a prototype for the OP swab sampling tasks.

Keywords: Robotic vision and image processing, Artificial intelligence in robotics
Abstract: With the increasing surgical need in our aging society, there is a lack of experienced surgical assistants, such as scrub nurses. To facilitate the training of junior scrub nurses and to reduce human errors, e.g., missing surgical items, we develop a speech-image based multimodal AI framework to assist scrub nurses in the operating room. The proposed framework allows real-time instrument type identification and instance detection, which enables junior scrub nurses to become more familiar with the surgical instruments and guides them throughout the surgical procedure. We construct an ex-vivo video-assisted thorascopic surgery dataset and benchmark it on common object detection models, reaching an average precision of 98.5% and an average recall of 98.9% on the state-of-the-art YOLO-v7. Additionally, we implement an oriented bounding box version of YOLO-v7 to address the undesired bounding box suppression in instrument crossing over. By achieving an average precision of 95.6% and an average recall of 97.4%, we improve the average recall by up to 9.2% compared to the previous oriented bounding box version of YOLO-v5. To minimize distraction during surgery, we adopt a deep learningbased automatic speech recognition model to allow surgeons to concentrate on the procedure. Our physical demonstration substantiates the feasibility of the proposed framework in providing real-time guidance and assistance for scrub nurses.

Keywords: Medical robotics, biomedical and rehabilitation engineering
Abstract: This paper presents an analysis of the degrees of freedom, accuracy, and workspace requirements for arch nail implantation surgery, based on which a new spinal surgery robot with a hybrid 4-DOF mechanism is designed and analyzed, and the kinematic parameters are calibrated using the MDH modeling approach. Our design combines a low-degree-of-freedom mechanism with a gimbaled passive joint structure, resulting in smaller dimensions, lower costs, and better flexibility than conventional robots with 5-DOF or higher. As the robot uses a hybrid mechanism, we first analyzed its degrees of freedom and transformed it into a single open chain to obtain an equivalent mDH parameter table to simplify the kinematic problem, and optimized the kinematic parameters based on the Jacobian matrix. Finally, we introduce a fixed-pose inverse kinematic algorithm to analyze the robot's workspace and dynamic model. Using the degree of freedom analysis, kinematic analysis, workspace analysis, and kinematic parameter calibration methods, this paper presents an analysis method for a multi-directional hybrid robot system and demonstrates that the 4-DOF robot meets clinical requirements. These analysis methods play an integral role in the mechanism analysis, kinematic analysis, and kinematic parameter calibration of the multi-DOF hybrid robot system.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Smart sensors and actuators, Applications of robotics and biomimetics
Abstract: Robotic-assisted needle intervention has shown great prospects in minimally invasive surgery (MRI), where the needle insertion device (NID) is an important component. Some sophisticated intervention practices require the NID to have the characteristics of high resolution, enough insertion force, and compact structure. In this work, a NID driven by an inertial piezoelectric actuator is developed and validated by experiments. The actuator consists of a triangular configured compliant driving mechanism and a piezoelectric stack installed inside to meet the requirements. A prototype with a compact structure of 86 × 104 × 41 mm^3 is fabricated to test the performance of the proposed NID. Both the no-load test and phantom test are carried out. Experimental results show that the developed NID reaches a resolution of 4.7 nm (fine positioning mode), a maximum velocity of 6.89 mm/s, and a maximum insertion force of 2.4 N. The resolution and structure volume are superior to the traditional NID, and the insertion force is higher than the same type of NID. A potential application of the proposed NID is high-precision needle intervention with limited working space, such as MRI-based needle biopsy.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Autonomous mobile robots and manipulators
Abstract: A surgical robotic system used for minimally invasive surgery (MIS) usually includes a patient-side cart, exchangeable instruments and a few accessories. During the preoperative preparation, the patient-side cart is usually manually driven by a nurse to the desired location to connect to the trocar/sheath for docking. Aiming at improving the efficiency of manual docking, this paper, to the best of the authors’ knowledge, for the first time, proposes an automatic parking solution for the patient-side cart based on both a designed headband marker that is installed on the cart and a closed-loop pose controller that utilizes the pose feedback from the headband marker under an external monocular camera. The experimental results show that the average position and orientation estimation errors for the designed headband marker at a distance (from the camera to the marker) of 1.5 m ~ 6 m are 2.21 mm and 1.08°, respectively. The final position and orientation errors of the proposed closed-loop pose controller are 3.91 mm and 1.24°, respectively, which is considered sufficient for automatic parking of the patient-side cart.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Soft robotics and liquid-metal robotics
Abstract: Nasotracheal intubation is a crucial emergency procedure that can improve the survival rate of injured patients. Compared to orotracheal intubation, nasotracheal intubation offers greater stability and comfort. However, due to the large curvature of the cavity, the procedure is highly challenging and requires skilled medical staff, which can be a shortage in emergencies and delay patient rescue. To solve this problem, a flexible and portable robot for nasotracheal intubation has been developed in this paper. The robot incorporates a continuum catheter, allowing the shape of the endotracheal tube's distal section to be controllable and making it easier to pass through the nasal cavity into the trachea. An endoscope is integrated into the distal section of the catheter to make the intubation process more intuitive, allowing doctors to observe the injuries through endoscopic images and adjust the treatment plan. Additionally, the robot is designed with a quick-detachable structure, enabling aseptic and low-cost intubation procedures. The structure and circuitry of the robot are designed for portability, with a smaller size and lower power consumption, which can work for a long time just powered by batteries. Finally, experiments on a mannequin validate that the robotic system effectively reduces the difficulty of tracheal intubation, enabling even non-medically trained operators to complete the procedure with the assistance of the robot.

Keywords: Applications of robotics and biomimetics, Human-robot interaction, Artificial intelligence in robotics
Abstract: This paper introduces a modular robotic system designed to provide targeted illumination for various work tasks by dynamically adjusting a spotlight's orientation. A distinguishing feature of our system is its compatibility with MATLAB via a direct interface, allowing users to effortlessly implement and test their own algorithms for object detection and tracking. By combining the capabilities of an actuator unit, spotlight, and camera, the system is able to recognize and track different objects, providing precise lighting tailored to various applications. A comprehensive exploration of the underlying geometric relationships, control mechanism as well as the object detection and tracking methodologies is performed. An experiment was conducted to provide insights into the system's real-world performance. The results demonstrate a high degree of reliability, with tracking accuracy predominantly constrained by the neural network. Our findings highlight the system's potential as a versatile tool, particularly beneficial for researchers and developers aiming to experiment with and enhance object recognition and tracking algorithms.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics, Soft robotics and liquid-metal robotics
Abstract: In our previous research we have developed a multi-legged robot that can climb uneven or warped walls and large pipes. The robot can adapt its body to these environments using passive mechanisms. In this paper, we improve the previous robot to walk on the horizontal ground and travel from the ground to the wall. We developed a new robot that has stretchable trunk, and we conducted experiments to demonstrate the effectiveness of the developed robot.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Multi-sensor data fusion and sensor networks, Artificial intelligence in robotics
Abstract: The very often occurrence of foot slippage imposes a significant challenge on the locomotion of multi-legged robots, and slip detection is the prerequisite to prevent robots from slippage. Slip detection methods based on additive sensors are susceptible to noise and the sensors are expensive and easily damaged, and the existing methods using proprioceptive information fail to distinguish which feet are slipping given the robot’s state in case multiple legs slip at the same time, due to strong inter-dependence among the body and multiple branches of leg mechanisms. Deep learning can establish the relationship between input and output using a large number of samples, avoiding the complex process of manual feature extraction and giving hope for a solution to this difficulty. In this paper, we present a deep learning-based approach to slip detection in quadruped robots. Our slip detector uses a convolutional neural network to fuse information acquired from proprioceptive sensors to infer the slip state to provide a priori knowledge to the controller. Simulation results indicate the effectiveness of the proposed approach in achieving accurate and immediate slip detection of each foot for agile locomotion without prior knowledge about the contact state in multiple scenarios, and the learning accuracy is greater than 95% in every scenario.

Keywords: Human-robot interaction, Applications of robotics and biomimetics, Multi-sensor data fusion and sensor networks
Abstract: Skill transfer from a human hand to a robot hand is extremely challenging. Currently, researchers typically use mapping of human hand postures onto the robot hand with the help of a data glove. This approach to skill transfer has yielded impressive results for grasping applications, but is difficult to implement for fine manipulation tasks. Humans use their sense of touch to handle various objects to perform fine manipulation tasks such as stitching, calligraphy, carving, and painting. Consequently, it is also important to measure the tactile data exerted by the human hand on the object while performing fine manipulation tasks. Previously we presented FingerTac - a wearable tactile sensor that can be worn interchangeably on human hands as well as on robot hands. FingerTac is capable of measuring 3-axis tactile data distributed at 20 sensing points around the user’s fingertip. In this paper we present ExoGlove (see Fig. 1), which is a combination of a mechanical exoskeleton and FingerTac. The mechanical joints of the ExoGlove are embedded with magnetic encoders which allow precise monitoring of joint angles between the human hand finger segments. As the ExoGlove is integrated with the FingerTac it allows roboticists to not only map the posture of the human hand to the robot hand but to also map the tactile data from the human hand while performing fine manipulation tasks. The sensor is able to measure angles with a mean absolute error of 2.76 degrees.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Multi-sensor data fusion and sensor networks
Abstract: The goal of this work is to identify a reliable and low-cost set of electronics that can be used to implement lower limb exoskeletons. This system can be programmed to perform a variety of control algorithms and thus be used for various human-assistive purposes. The proposed system uses Controller Area Network (CAN) protocol for motor communication and Serial Peripheral Interface (SPI) for sensor data collection. An Ethernet link is used to send run-time data from the microcontroller unit (MCU) to a Wi-Fi access point that connects to a remote base station for data logging. Timing analysis is performed for the communication protocols used and statistical measures are calculated to determine a reliable lower bound for operational bandwidth of the system.

Keywords: Medical robotics, biomedical and rehabilitation engineering
Abstract: Self-balancing exoskeleton robots enable hands- free walking for the wearer, and the problem of its stable walking has been a challenge. In this paper, to achieve compliant and stable walking of a self-balancing exoskeleton robot and reduce the shock between the robot and the ground, we propose a new fractional kelvin voigt viscoelastic controller (FKVC). The proposed FKVC incorporates the advantages of fractional order and the kelvin voigt viscoelastic to endow the self- balancing exoskeleton with stable and dynamic walking capa- bilities. Simulation results show that the actual CoM position of the exoskeleton well tracks the reference CoM position, and the actual ZMP converges to the reference ZMP position. The ZMP is always within the effective support region of the biped. Therefore, we experimentally proved the practicality and effectiveness of the FKVC method, and the self-balancing exoskeleton successfully realizes stable dynamic walking.

Keywords: Medical robotics, biomedical and rehabilitation engineering
Abstract: Lower back pain (LBP) is a common problem addressed among adults working in physically demanding environments. As a countermeasure, we have developed the Muscle Suit series, an assistive device designed for waist support. In 2019, we launched the mass-produced model “Every”, making the product more accessible by reducing the price from over 500,000 yen to 100,000-yen range, and more than 20,000 units have been shipped to date. However, many users have requested improvements in terms of size, weight, and ease of walking. This report introduces the major points of improvement we have made to Every to meet these user’s demands.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Multi-robot systems, swarm robots, and collaborative robots, Human-robot interaction
Abstract: To mitigate the risk of musculoskeletal disorders and injuries at the back, occupational back-support exoskeletons are being introduced in the workplace because of their ability to reduce workers’ lumbar loading during the execution of demanding activities. Nevertheless, back-support exoskeletons’ assistance has not been developed extensively for activities different from lifting. This work presents a control strategy for an active back-support exoskeleton specifically tailored for assisting carrying task and analyzes the physiological effects on the user. The control strategy modulates the exoskeleton torques to assist the back without hindering the user’s legs’ during gait. A gait phase detection framework that uses sensorized insoles is employed to control the modulation of the assistive torques. The strategy is assessed through an experimental test on nine subjects. Participants performed a simplified carrying task without and with the exoskeleton. The experimental evaluation found a statistically significant increase in the normalized heart rate when using the exoskeleton compared to not using it. No statistically significant changes were found in the mental effort between the conditions with and without the exoskeleton aid.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Artificial intelligence in robotics, Multi-sensor data fusion and sensor networks
Abstract: To improve the human-robot matching performance of exoskeleton robots, a fine sensing method oflower limb motion patterns based on a support vector machine optimized by a sparrow search algorithm and electromyographic signals (EMG) is proposed. Firstly, the EMG signals of four kinds of lower limb movements, namely, stair climbing, stair descending, standing and sitting, and walking, are collected; then, the collected EMG signals are pre-processed and feature extracted to obtain the eigenvalues of the EMG signals in the time domain and the frequency domain; finally, the optimized Support Vector Machine based on the sparrow searching algorithm (SSA-SVM) is utilized to complete the identification of lower limb movement patterns, which is compared to the classification results of the traditional SupportVector Machines (SVMs) and Temporal Convolutional Networks (TCNs). The classification results show that the recognition accuracy of the Support Vector Machine (SSA-SVM) classification model optimized based on the Sparrow Search Algorithm is 98.9%, the recognition accuracy of the SVM classification model is 90.7%, and the recognition accuracy ofthe TCN classification model is 89.4%, and the SSA-SVM classification model has the highest classification accuracy.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Applications of robotics and biomimetics, Human-robot interaction
Abstract: The number of upper limb amputees all over the world is high and continue to increase. They need a prosthetic hand to have a more comfortably life. Grasping a thing properly with a prosthetic hand can smooth the movement of an upper limb amputee and expand the range of possible activities. However, few prosthetic hands can automatically maintain a stable grasping posture, and most of them require the user to pay attention to doing it. In this study, we developed a new robotic prosthetic hand that can automatically maintain a stable grasping posture by using a wrist rotation mechanism, focusing on the center position of the grasping force from pressure sensors attached to each finger. This hand rotates its wrist according to the center position of the grasping force, which depends on the changing center of gravity position of the object grasped by the hand. From the results, it is confirmed that the grasp is maintained stably

Keywords: Soft robotics and liquid-metal robotics, Human-robot interaction, Artificial intelligence in robotics
Abstract: In this paper, the concept of ``wearable" and ``flexible" is combined to design and develop a wearable flexible robot arm, which can realize multi-person collaboration and human-computer interaction by adapting the corresponding end-effector. The main work includes: design and research on the structure and manufacturing process of wearable flexible robotic arm, mainly designing a casting molding process based on spiral compression scheme, and completing the design and manufacturing process of flexible robotic arm; A closed-loop control system for the flexible manipulator was built, and a control strategy based on the combination of iterative learning control and convolutional neural network was proposed. Through model training and simulation, it was proved that the proposed method can effectively improve the control accuracy of the end of the manipulator. An experimental test platform for the robotic arm was built, and related experiments were conducted. The experimental results showed that the wearable flexible robotic arm developed in this paper had a certain degree of completion, its motion effect and control accuracy reached the design expectations, and it had certain application prospects.

Keywords: Artificial intelligence in robotics, Medical robotics, biomedical and rehabilitation engineering
Abstract: Estimation of hip joint torque is of great significance for exoskeleton assist torque planning. However, traditional single neural network models are difficult to reliably estimate human joint torque and the dynamic models based on physical theorem are limited by the measurement technology of ground reaction force. Therefore, this paper proposes a parallel fusion neural dynamic model that incorporates LSTM, NTM, and Newton-Euler dynamical equation. The model only needs human kinematic parameters as inputs to complete the estimation of human hip joint torque. To evaluate the estimation performance, this paper introduces relative accuracy as an evaluation standard. The experimental result shows that the estimation performance of the fusion model is greatly improved compared with the traditional single neural network models. The fusion model proposed in this study can be used to estimate the torque of the hip joint. It can be integrated into the exoskeleton control system and used as the basis for planning the assisting torque of the exoskeleton.

Keywords: Artificial intelligence in robotics, Medical robotics, biomedical and rehabilitation engineering
Abstract: Abstract—To improve the control of exoskeleton robot for gait assistance, a gait phase recognition method for exoskeleton assistance based on three inertial sensors is proposed, which includes motion state recognition based on support vector machine (SVM), gait phase recognition relying on long short-term memory (LSTM), and a rule-based method for labeling lower limb gait event. Firstly, we used SVM to identify six motion states: standing, level walking, turning left, turning right, left steering, and right steering. Then, an LSTM model was employed to identify the six phases of the lower limb gait during level walking. The experiments demonstrate that this method achieves accuracy rates of 99.65% and 96.81% for the wearer's motion state recognition and gait phase recognition, respectively, with corresponding F1 scores of 0.9957 and 0.9683.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Artificial intelligence in robotics, Applications of robotics and biomimetics
Abstract: This paper presents a general estimation method of deformation for the Self-balancing lower limb exoskeleton (SBLLE). In particular, we propose a Bi-LSTM deformation estimator (BLDE) to predict and compensate for the deformation of SBLLE based on the current force and torque data measured by force/torque (F/T) sensors. First, we choose four movements including squatting down and up, waist twisting, left foot lifting, and right foot lifting to mimic the constituent action of walking motion. The deformation data set is obtained through the motion capture analysis system, and the relative F/T data set is obtained by the F/T sensors embedded in the feet of SBLLE. Second, the Bi-LSTM network is trained to learn the relationship between the deformation and F/T and verified on the test set. After that, BLDE is added to the control system of SBLLE to estimate the deformation. Finally, four same movements and the walking experiment are conducted on the exoskeleton AutoLEE-G2 with BLDE. The experimental results have proven that BLDE can predict and compensate for deformation by only using F/T sensors.

Keywords: Human-robot interaction, Artificial intelligence in robotics, Multi-robot systems, swarm robots, and collaborative robots
Abstract: Safe and effective control of robots to assist humans is complex, especially when it involves close physical interactions with humans. This work presents a hierarchical framework to build robot controllers to assist humans in walking under frailty constraints. We propose to train an RL-based human policy in simulation, to model and synthesize human walking behaviours under external assistance and frailty constraints. It provides a testbed of safe and scalable interactions with humans, and thus allows for iterative fine-tuning of robot behaviours to offer physical assistance robustly. The efficacy of the proposed framework has been evaluated on a dual-arm assistive robot. Experimental results show that the learned walking policy enables humans to leverage random external assistance to generate and stabilize walking motions under frailty constraints. We also demonstrate that the robot controller obtained from interacting with the human is more effective and robust to assist the frail human in walking.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Human-robot interaction, Applications of robotics and biomimetics
Abstract: The human-in-the-loop optimization strategy has successfully improved the performance of the exoskeleton, but only considering the effect of physiological parameters, such as individual metabolic cost and joint torque differences, are considered. From a psychological perspective, emotion is also a significant factor affecting human gait character, especially during walking, where there are differences in the biological hip torque under different emotions. Inspired by this, we developed an individualization assistance strategy for hip extension based on Emotion-in-the-Loop (EIL) to improve human-exosuit interaction using emotional feedback. First, we classified four typical emotions based on the Support Vector Machine (SVM), the average accuracy of which reached 92.8%. After that, measuring the biological hip moments under different emotions through a motion capture experiment and the individualization of assistance strategy for hip extension assistance was designed based on the result. In the metabolism experiments, the soft exosuit with the EIL method can significantly further reduce metabolic consumption by about 4.5%, 6.7%, and 3.7% under the emotion of excitement, depression, and fear, respectively, in addition to original metabolic consumption reduction of 7.6%, which demonstrated that applying different assistance forces considering emotional changes can considerably reduce metabolic consumption. In the future, this method can be additionally used on different wearable robots to improve the performance of human-robot interaction.

Keywords: Medical robotics, biomedical and rehabilitation engineering, New theory and technology in robotics and biomimetics, Soft robotics and liquid-metal robotics
Abstract: Robotic hand exoskeletons can assist patients with hand disabilities and the elderly to perform their activities of daily living (ADL). However, designing a segmented variable curvature flexible hand exoskeleton robot is still a great challenge. In this paper, a wearable flexible finger exoskeleton based on a double-layer parallel flat spring structure is first proposed by designing different widths of the segments of the fixed-length flat springs, which can achieve different curvatures. Second, a variable curvature kinematic model based on the relative displacement is established and analyzed. Third, a variable curvature kinematic model based on the push force is established and analyzed considering the effects of friction and external forces. The curvature value of each segment is obtained by recursively calculating the forces and moments of each segment. Finally, the accuracy of the two kinematic models is verified by experiments. The experimental results show that the variable curvature motion model with friction is more accurate than that without friction. Moreover, wearing experiments show that the flexible hand exoskeleton has the potential for ADL assistance application.

Keywords: Smart sensors and actuators, Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics
Abstract: Magnetorheological (MR) fluids-based devices such as MR dampers and brakes can output a continuously controllable damping force via modulating the magnitude of the applied magnetic field. Owing to low energy consumption and fast response, recently vast MR dampers and brakes are widely studied and explored for lower limb prostheses and exoskeletons. This paper proposed a novel MR brake with adjustable interdigitated comb for lower limb prostheses. Through actively adjusting the area the of the comb engagement, the maximum viscous and field-induced damping torque can be increased or decreased. Because of that, the dynamic adjustment interval of the output damping torque can be increased to meet the function requirements of the prosthetic knee in different walking modes. Magnetic analysis via ANSYS is performed to check the magnetic density of the MR fluid under different excitations. Mathematical models of the damping torque are built. Finally, a prototype of the novel MR brake is fabricated.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Human-robot interaction, Applications of robotics and biomimetics
Abstract: In response to the deficiency of traditional rigid-driven exoskeletons, this paper presents the mechanical design and control strategy of a compact variable stiffness joint mechanism for modular lower limb assistive exoskeletons. The stiffness variation is achieved by modifying the lever arm principle, while allowing for adjustable configurations to cater to different stiffness requirements. The range of stiffness variation of the proposed joint is first mathematically modeled, and its stiffness characteristics, as well as the effects of different configurations, are then analyzed. In accordance to the mechanism design of the joint, a control strategy for a modular compliant lower limb assistive exoskeleton is proposed. The stiffness adjustment and output torque requirements are analyzed and modeled and a solution to control variables are derived for the 40% assistance case. The case study results indicate a relatively low coupling between the stiffness and torque of the mechanism.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Applications of robotics and biomimetics
Abstract: Stroke has become one of the main causes of disability and death in the elderly population. A number of stroke survivors suffers from motor dysfunction, reducing their daily quality of life. Early intervention can effectively improve rehabilitative effectiveness. Therefore, it is necessary for patients to start rehabilitation training when they are in bed. In this paper, a bedside rehabilitation robot is proposed to assist patient’s lower limb rehabilitation exercise training from bedridden stage. A motor drives a double-link mechanism to achieve linear reciprocating motion for leg extension and retracting training. Two double-link mechanisms allow stroke patients to perform different training modes. A hill model of lower leg muscle was constructed to simulate the motor workload during linear reciprocating motion of patient lower limb. The preliminary experiments demonstrated the feasibility of the bedside rehabilitation robot for lower limb exercise training of stroke survivors.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Human-robot interaction
Abstract: Rehabilitation robots play an important role in the motor function rehabilitation for stroke survivors with hemiplegia. However, the rehabilitation effect of current robots is still limited partly because a single training of motor function can be strongly affected by the decreased blood supply function of the bedridden patients. This work proposed an approach to study the coupling relationship between the motor and blood supply functions by combining the synchronously recorded EEG and cerebral blood oxygen information, where the cerebrations in different movement paradigms were analyzed from an aspect of “functional coupling”. The results show that the information of oxyhemoglobin concentration change (ΔHbO) can effectively indicate the cortex activation, and a stronger blood supply is needed in cortexes to perform body movements. The correlations within motor cortexes are significantly stronger than the ones between motor and prefrontal cortexes, and a higher resistance level of extremity training will cause stronger correlations. Calculation of transfer entropy (TE) shows that there exists a bidirectional information transmission between the electrophysiological and blood supply signals, and more information is transmitted always in the direction from ΔHbO to EEG than in the opposite direction. The information transmission or the coupling relationship can be significantly enhanced by extremity movements, and large TE values are achieved in the Theta, Beta and Gamma frequency bands of EEG that correspond to motor functions. This work has demonstrated the functional coupling between nerve and blood microcirculation, which would provide a technical guidance to improve the rehabilitation effect for current robot systems and have great application potentials.

Keywords: Multi-sensor data fusion and sensor networks, Smart sensors and actuators, Human-robot interaction
Abstract: Gait recognition plays a significant role in the realm of wearable robotics, particularly in facilitating the development of sensors for gait analysis in rehabilitation. In this study, we proposed a novel optic-based sensor for gait recognition and employed four neural network models to extract both temporal and spatial information. Through a comparative analysis of these neural network models across the gait cycle, we investigate the implications of different sensor placement locations and the contributions made by diverse types of information in the context of gait analysis.Gait recognition plays a significant role in the realm of wearable robotics, particularly in facilitating the development of sensors for gait analysis in rehabilitation. In this study, we proposed a novel optic-based sensor for gait recognition and employed four neural network models to extract both temporal and spatial information. Through a comparative analysis of these neural network models across the gait cycle, we investigate the implications of different sensor placement locations and the contributions made by diverse types of information in the context of gait analysis.

Keywords: Applications of robotics and biomimetics, Human-robot interaction
Abstract: Lower limb exoskeletons for gait rehabilitation have been widely studied, including rigid and flexible exoskeletons. In order to improve the human-robot coordination during exoskeleton wearing, avoid the constraints on joints caused by rigid structures, this paper designed a lightweight lower limb exoskeleton with compliant knee joints. By adopting a lightweight design, the main body is composed of carbon fiber and aluminum alloy, resulting in a total weight of only 7KG. Different from the current knee joint cable-driven structure that install the joint motor on the back, we install the knee joint motor on the thigh and use two completely parallel Bowden cables to transmit force. This structure can reduce the mass and rotational inertia at the knee joint, improve the wearable characteristics of the exoskeleton, and reduce the additional frictional losses caused by long-distance force transmission of the Bowden cables. Meanwhile, the exoskeleton adopts a modular design, which facilitates its application in different assistive modes. Finally, through the joint encoders and motor PD control to track the joint trajectory, the measurement results show that the error for the hip and knee joints is within 2° and 3.5° respectively.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Human-robot interaction, Applications of robotics and biomimetics
Abstract: 本文介绍了一种创新设计的柔性 基于柔性螺钉的手部康复外骨骼 机制。提议的柔性手的手指 由弹性结构组成，通过大驱使 FSM的变形，可实现双向 主动弯曲。基于重新设计的FSM，这只手 外骨骼可以做得紧凑、轻便且易于 控制。此外，外骨骼的四个手指 独立驱动，使多元化 康复训练。在这项研究中，我们介绍了 柔性手指的机械结构，以及 柔性弯曲角之间的关系 手指和FSM的线圈数量被制定， 并制定误差补偿模型 实验。最后，手部外骨骼的原型 介绍了屈曲伸展和 手部外骨骼的抓握力是初步的 通过实验评估。

Keywords: New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: Electroencephalography (EEG), as a non-invasive and convenient method for implementing Brain-Computer Interface (BCI), has been widely used in clinical and research fields. EEG data often requires the acquisition of dozens or even hundreds of channels. Channel selection can reduce irrelevant and redundant channels, improve computational efficiency, and enhance the quality of EEG signals. This study introduces a filter method for channel selection based on Pearson correlation coefficient (PCC) with the candidate channel and employs topographic maps of EEG channel scores, derived from data collected across all subjects, to visualize the spatial distribution of channels selected by different methods. In addition, a generalized channel selection algorithm is proposed to determine consistent channels across all subjects in the experimental group. The effectiveness of the proposed method was evaluated on two steady-state visual evoked potential (SSVEP) datasets, and the results indicated that this method exhibits superior performance compared to both the all-channel method and other channel selection methods. And the application of the generalized channel algorithm has further improved the classification performance. This study uses selected generalized channels applied to new subjects with low BCI performance, yielding a significant improvement. The selected channels have a wide range of applicability, helping to simplify EEG acquisition and improve EEG data quality.

Keywords: Robotic vision and image processing, Human-robot interaction, Artificial intelligence in robotics
Abstract: In human-robot interaction (HRI), human pose estimation is a necessary technology for the robot to perceive the dynamic environment and make interactive actions. Recently, graph convolutional networks (GCNs) have been increasingly used for 2D to 3D pose estimation tasks since the skeleton topologies can be viewed as graph structures. In this paper, we propose a novel graph convolutional network architecture, Multi-scale Multi-branch Fusion Graph Convolutional Networks (MSMB-GCN), for 3D Human Pose Estimation(3D HPE) task. The proposed model consists of multiple GCN blocks with a multi-branch architecture. This multi-branch architecture enables the model to get multi-scale features for human skeletal representations. The group of GCN blocks, which has strong multi-level feature extraction capabilities, allows the model to learn global and local features, lower-level and higher-level features. Experiment results on the HumanPose benchmark demonstrate that our model outperforms the state-of-the-art and ablation studies validate the effectiveness of our approach.

Keywords: Human-robot interaction, Robotics in intelligent manufacturing, Applications of robotics and biomimetics
Abstract: Industrial exoskeletons are becoming more important to prevent workers from musculoskeletal disorders when performing manual material handling. Since the exoskeletons are task dependent, there is a versatility of use in the performance according to the type of movement that the user executes. The User Command Interface is a wearable device for industrial exoskeletons. It is an adaptable setup system between users and occupational wearable robots. The main purpose is to give access to users into some domains of the exoskeleton configuration system. Areas such as secure user identification, parameter configuration, and signal visualization are well structured into the interface. Shoulder-sideWINDER is an active exoskeleton for shoulder support developed in collaboration between XoLab and INAIL (Italian Worker's Compensation Authority). This exoskeleton needs a guided calibration process before starting the operation. In this paper, we present the evaluation of two novel features of the UCI: (a) Tutorial and (b) Audio Notifications for calibration. These new features are integrated into the upper-limb industrial exoskeleton Shoulder-sideWINDER. Participants performed experiments for assessing the interface understandability and usability of tutorial and sound notifications by wearing the Shoulder-sideWINDER exoskeleton and the User Command Interface. The results presented in this paper show that the User Command Interface is a highly intuitive control device to guide the calibration sequence of the exoskeleton.

Keywords: Applications of robotics and biomimetics, Medical robotics, biomedical and rehabilitation engineering, New theory and technology in robotics and biomimetics
Abstract: Comfort in prosthetic sockets remains a significant challenge for many amputees, particularly for above-knee amputees bearing substantial weight on their soft tissue [1], [2]. The predominant source of discomfort often originates from swelling of the residual limb, a factor traditional prosthetic sockets fail to accommodate [3]. While Osseointegration surgery, a unique method to connect prosthetics and residual limbs while removing the socket, is recognized as a superior solution for enhancing comfort and quality of life, its accessibility remains limited [4]. Hence, the motivation for this study is to explore the development of an automatically actuating and adjustable socket. This research aims to design and develop a prosthetic limb socket offering enhanced comfort, flexibility, and user control i.e. a robotic socket (RS). A comprehensive literature review was conducted to assess existing interactive sockets and sensory systems, focusing on their performance and specifications. The methodology employed in this study (Phase 1) involved the design and testing of small-scale circuits which will later (Phase 2) evolve into a more complex and up-scaled system. The prototype developed in this study features a plastic bottle equipped with a pressure valve, simulating residual limb swelling by modulating the internal pressure. A 3D-printed button and locking wire serve as a rigid socket, preventing unchecked bottle expansion. A motor integrated into the system modulates hoop tension on the conceptual socket, ensuring user-specific pressure regulation. The system was developed using a Raspberry Pi to manage motor fluctuations based on user-defined set points and bottle pressure readings. Preliminary testing demonstrated promising results, with the motor effectively adjusting

Keywords: Medical robotics, biomedical and rehabilitation engineering, Applications of robotics and biomimetics, Human-robot interaction
Abstract: This paper presents the design and experimental validation of a 2-DoF elbow exoskeleton with a 2-DoF spherical scissor mechanism (SSME) to assist patients with upper limb impairments in activity-of-daily-living (ADL) tasks. Since the forearm pronation/supination was missing in most portable exoskeletons and the existing design with forearm rotation were too bulky, we propose a 2-DoF elbow exoskeleton with a spherical mechanism for elbow and forearm assistance. Benefiting from the spherical scissor and cable-driven mechanisms, the SSME can assist in elbow flexion/extension and forearm pronation/supination, up to 97.467% ADL workspace, with a mechanical design as light as 0.543 kg, practical for daily use. Besides, the proposed sensing method and kinematic model for the spherical scissor mechanism can measure human joint motions, solving the sensing challenge for forearm motion in most existing upper limb exoskeletons. Furthermore, with inverse kinematics and robot-joint-space impedance control, the SSME can independently control the elbow and forearm angle with minimal influence on another, regardless of its coupled nature. Moreover, our experiments have validated the proposed exoskeleton design in terms of the force control performance of actuators, robot joint impedance control response, and robot assistance in an ADL task.

Keywords: Medical robotics, biomedical and rehabilitation engineering
Abstract: Passive prostheses lack the ability to generate power for the body, potentially increasing the user's energy expenditure. However, a powered prosthesis can provide power, but its volume and weight of energy limit the range of motion of the prosthesis. This paper presents a proposal for a powered ankle-foot prosthesis using a direct-drive electro-hydraulic system and a three-chamber hydraulic cylinder that mimics the human lower limb. An iterative learning control method combined with PID control is used to control the energy saving ankle-foot prosthesis with a three-chamber hydraulic cylinder. The direct-drive electro-hydraulic system used for the three-chamber hydraulic cylinder in this paper is designed, mathematically modelled and simulated to ensure its validity. The simulation results show that the energy saving ankle-foot prosthesis with a three-chamber hydraulic cylinder can reduce energy consumption and improve range compared to the conventional ankle-foot prosthesis with a two-chamber hydraulic cylinder. Furthermore, the combined implementation of the iterative learning control method and PID control proves to be effective in reducing errors.

Keywords: Artificial intelligence in robotics
Abstract: 在不同材料和工艺参数条件的两相流实验中，采集到的相似度低、数量少的图像数据集对于普通深度学习算法来说难以实现流型的高精度识别。 由于全局特征的提取能力低。在本文中，我们提出了一种新的深度学习算法，通过CNN增强Swin-T网络，该算法结合了Swin-T网络与动态区域感知卷积的优点。新算法保留了窗口多头自注意力机制，并增加了自注意力调整模块，增强了图像特征的提取和网络的收敛速度。它显著提高了清晰和模糊图像中不同流型的识别精度。CNN的增强网络Swin-T对图像数据集分类具有较高的相似度和少量

Keywords: Multi-sensor data fusion and sensor networks, Artificial intelligence in robotics, Robotic vision and image processing
Abstract: Multi-object tracking (MOT) has important applications in monitoring, logistics, and other fields. This paper develops a real-time multi-object tracking and prediction system in rugged environments. A 3D object detection algorithm based on Lidar-camera fusion is designed to detect the target objects. Based on the Hungarian algorithm, this paper designs a 3D multi-object tracking algorithm with an adaptive threshold to realize the stable matching and tracking of the objects. We combine Memory Augmented Neural Networks (MANN) and Kalman filter to achieve 3D trajectory prediction on rugged terrains. Besides, we realize a new dynamic SLAM by using the results of multi-object tracking to remove dynamic points for better SLAM performance and static map. To verify the effectiveness of the proposed multi-object tracking and prediction system, several simulations and physical experiments are conducted. The results show that the proposed system can track dynamic objects and provide future trajectory and a more clean static map in real-time.

Keywords: Robotic vision and image processing, Artificial intelligence in robotics
Abstract: Low-light image enhancement (LIE) aims to enhance image contrast and recover fine details for images captured in low-light conditions. However, the limitations of using a single image and manually defined priors lead to inadequate information and limited adaptability, resulting in the failure to reveal image details effectively. To tackle this problem, we propose an unsupervised dual information flow LIE model that learns adaptive priors from low-light image pairs. The model is based on the assumption that priors can be learned from pairs of images. Moreover, a simple self-supervised model is designed to perform feature processing on the original input and further processing on the image to avoid the suboptimal limitations of the model. As a result, our model exhibits superior performance on the LIE task compared to other algorithms. As a result, our model exhibits improved network adaptability and superior performance in the LIE task compared to other algorithms.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots, Robotics in intelligent manufacturing, Artificial intelligence in robotics
Abstract: Small robots encounter considerable difficulties in learning effective motions on complex terrains owing to their underactuated nature and nonlinear dynamics. In this paper, we present a novel approach for robot motion generation that implements reinforcement learning, based on simplified exploration of the robot's action and time slice conduction. Our approach controls the robot's actions using normalized signals and hierarchical mappings on mathematical space, which facilitates the learning process. We execute action in the timeslice to make efficient interaction with the environment. The effectiveness of our methodology is evaluated across a diverse range of simulated terrain scenarios, supplemented by physics validation. Our results show that our approach performs effective on complex terrains that are designed for small-sized robots.

Keywords: Artificial intelligence in robotics, Autonomous mobile robots and manipulators, Robotic vision and image processing
Abstract: Currently, the majority of research on humanoid robot basketball shooting focuses on traditional control methods. However, these methods primarily rely on human-robot interaction and fixed shooting patterns to control the robot's shooting actions, resulting in limited autonomy for the robot. They often require extensive manual design and coding operations, and face challenges in adapting to different shooting scenarios. To address these problems, this paper applies deep reinforcement learning to the basketball shooting task for a humanoid robot. The task environment is based on the basketball shooting competition defined in the FIRA HuroCup. This paper uses the Double DQN algorithm to train the humanoid robot to master end-to-end basketball shooting skills, specifically: The robot takes RGB images captured by its own head camera as input, then decides to take one of three discrete actions, including turning left, turning right, and shooting. In the experimental section, we validate the effectiveness of our approach and conduct an analysis and discussion on the setup of important parameters that influence the experimental results.

Keywords: Space robots, aerial robots, and underwater robots, Human-robot interaction, Robotics in intelligent manufacturing
Abstract: Micro/low gravity is one of the most prominent features of the outer space environment, and it significantly alters the force state and dynamics of spacecraft or astronauts compared to the Earth's gravitational environment. It is crucial to simulate the micro/low gravity environment on the ground for astronaut training or spacecraft testing. The suspension method utilizes a pulley and sling mechanism to create a micro-low gravity environment. This method counteracts the gravitational force exerted by the object based on rope tension. The simulation effect greatly depends on the accuracy of the horizontal following system, which serves as the central subsystem of the suspension device. In this paper, we propose a dual-arm following system to solve the issue of horizontal following for self-momentum targets. In addition, we conduct research on adaptive inhibition technology for flexible rope swing, and coupling control between a robotic arm and a crane. Physical experiments are conducted on the robotic system to verify the effectiveness of the proposed approach.

Keywords: Applications of robotics and biomimetics, Space robots, aerial robots, and underwater robots
Abstract: Lugged wheels are widely used to increase wheels' traction on sandy terrain such as planets. It is essential to model the interaction between lugged wheels and sand for predicting robot behavior, proper design, and control. In previous studies, wheels with an actively actuated lug were developed to achieve high-motion performance on sandy terrain. However, the lack of a motion prediction model limited the potential of robot control in these situations. This paper introduces a mathematical model of a wheel with an active lug based on terramechanics. Comparing the simulation and experiment results, the validity of the proposed model is shown.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, New theory and technology in robotics and biomimetics, Human-robot interaction
Abstract: Nowadays, several real-world tasks require adequate environment coverage for maintaining communication between multiple robots, for example, target search tasks, environmental monitoring, and post-disaster rescues. In this study, we look into a situation where there are a human operator and multiple robots, and we assume that each human or robot covers a certain range of areas. We want them to maximize their area of coverage collectively. Therefore, in this paper, we propose the Graph-Based Multi-Robot Coverage Positioning Method (GMC-Pos) to find strategic positions for robots that maximize the area coverage. Our novel approach consists of two main modules: graph generation and node selection. Firstly, graph generation represents the environment using a weighted connected graph. Then, we present a novel generalized graph-based distance and utilize it together with the graph degrees to be the conditions for node selection in a recursive manner. Our method is deployed in three environments with different settings. The results show that it outperforms the benchmark method by 15.13% to 24.88% regarding the area coverage percentage.

Keywords: Space robots, aerial robots, and underwater robots, Multi-robot systems, swarm robots, and collaborative robots, Applications of robotics and biomimetics
Abstract: Autonomous exploration and mapping of unknown subterranean areas using fleet of Micro Aerial Vehicles (MAVs) have achieved great interest in the past decade. The application of such research direction is extended beyond mining industry for example the exploration of potential lava tubes and caves on Mars using Martian MAVs. In this article we present a Collaborative Exploration and Mapping approach for exploration of large scale multi-branched virtual cave environment using multiple MAVs. An important aspect of such approach is establishing a decent collaboration with minimal information exchange among agents such that the exploration area can be maximized by preventing multiple agents covering same part of the environment. The collaborative exploration approach in this article is based on sharing local traversable frontier points as information among agents such that the each agent can identify the target areas in order to extend global map in an efficient manner. This work proposes a novel baseline collaborative exploration approach for optimal target area selection with a resource constrained MAVs in order to efficiently as well as rapidly explore subterranean environment. Extensive simulations are performed to validate the proposed collaborative strategy in order to establish a multi agent coordinated exploration-coverage mission.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Bio-inspired robots, e.g., climbing, creeping, and walking robots, New theory and technology in robotics and biomimetics
Abstract: Shepherding system is used one or more robots to guide an escaping target to a destination. It is said that this system can be applied not only to guide sheep flock, but also to various real-world applications such as extinguishing forest fires by guiding flames, guiding human. Therefore, a guidance system that can real-world conditions with a variety of disturbances is considered necessary. However, most of the existing guidance systems focus on how efficiently they can guide a target. On the other hand, there are also studies that aim for adaptive behavior by using methods such as machine learning to change the target's state. But these studies have not been able to guide targets that change their states multiple times or behave in unexpected ways. In this study, we propose an algorithm that can adaptively guide unknown targets that have not been modeled in advance. The target adaptation method is based on the social learning theory, which is an algorithm for social cognition, and uses imitation of robot behavior as personality imitation and independent learning by robots. This paper does not aim to optimally guide a target with a certain behavior. The goal is to achieve a certain level of performance in guiding a target whose behavior has not been modeled in advance. The results show that the parameters of the robot are tuned to the target by personality imitation and independent learning. In the future, we plan to improve the independent learning algorithm to achieve higher guidance performance and improve the independent learning algorithm to achieve higher guidance performance.

Keywords: Autonomous mobile robots and manipulators, New theory and technology in robotics and biomimetics, Multi-robot systems, swarm robots, and collaborative robots
Abstract: With its advantages of high efficiency and robustness, the multiple robots system has better application prospects in dynamic and complex environments, such as smart factories. However, when the number of robots increases, it is critical to efficiently coordinate congestion and jams among robots. To address this problem, we design the Double-Layer Multi-robot Path Planner to coordinate the robot conflicts. In the Collaboration Layer, we propose a Deep Multiple Dueling Network Agent( DMDNA) for node distribution based on the idea of forward node distribution to coordinate the path conflict problem of multiple robots. In DMDNA, we design the Multiple Dueling Network(MDN) model based on the DDQN network structure to address the high dimensional discrete action space problem in multiple robots. In addition, to overcome the characteristics of discrete and sparse rewards of reinforcement learning, we add the Hindsight Experience Replay(HER) experience replay strategy in the training of DMDNA. The experience of primary sequences is cut and reused to improve the utilization of samples. In the Motion Layer, we propose the Adaptive-DWA(A-DWA) algorithm, which adds the Target Function. According to the Target Function, the forward simulation time and weight of the evaluation function can be dynamically changed to improve the efficiency of local path planning and dynamic obstacle avoidance. The final experimental results show that the Double-Layer Multi-robot Path Planner based on DMDNA has better success rate and time cost than traditional reinforcement learning methods in solving multi-robot path planning problems. Relevant experimental code: https://github.com/WILL-ZHAO-1/DMDNA CODE.git.

Keywords: Artificial intelligence in robotics, Multi-robot systems, swarm robots, and collaborative robots, Robotics in intelligent manufacturing
Abstract: It is well known that it is difficult to have a reliable and robust framework to link multi-agent deep reinforcement learning algorithms with practical multi-robot applications. To fill this gap, we propose and build an open-source frame- work for multi-robot systems called MultiRoboLearn 1 . This framework builds a unified setup of simulation and real-world applications. It aims to provide standard, easy-to-use simulated scenarios that can also be easily deployed to real-world multi-robot environments. Also, the framework provides researchers with a benchmark system for comparing the performance of different reinforcement learning algorithms. We demonstrate the generality, scalability, and capability of the framework with two real-world scenarios 2 using different types of multi-agent deep reinforcement learning algorithms in discrete and continuous action spaces.

Keywords: Multi-robot systems, swarm robots, and collaborative robots
Abstract: In this paper, a decentralized relay-based approach (D-MRFTE) for unknown area exploration using a team of autonomous mobile robots is proposed under communication constraints. Using the relay robots, the multi-robot system forms a high-latency decentralized network with distributed copies of exploration information for which eventual consistency and completeness are ensured through meetups. The meetups act as a safety net and set a bound on latency by ensuring data transfer at periodic intervals whenever the multi-robot network gets fragmented. The information exchange related to the robot’s state and the ongoing exploration is facilitated by the relay robots. The robots use timestamps to assimilate the latest available information by using version vectors. To achieve a consistent state of explorer robots, the relays schedule meetups with other relays they come in contact with, creating a tightly-knit group. Our approach, under two communication models, i.e., Disk-based and Line-of-Sight-based, exhibits superior performance compared with two state-of-the-art algorithms in terms of completion time and distance traveled by the robot team. The simulations are conducted in a Player/Stage simulator with different robot team sizes.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Autonomous mobile robots and manipulators, Applications of robotics and biomimetics
Abstract: Applying Wireless Sensor Networks(WSNs) to coal mine safety monitoring is an effective means to achieve continuous monitoring of underground environmental parameters, but how to deploy them is a prerequisite for application. Firstly, a three-dimensional deployment strategy based on Confident Information Coverage (CIC) model suitable for narrow and long underground spaces is proposed. Secondly, an autonomous deployment method for WSNs nodes in a multi robots system using robots as node carriers is proposed, which effectively improves the deployment efficiency of nodes. Finally, the deployment strategy and method proposed in this paper were validated through simulation experiments. The experimental results showed that the proposed strategy can reduce the number of nodes, and the deployment method is also superior to existing methods.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Space robots, aerial robots, and underwater robots, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Modular robots are expected to be used in extreme environments owing to their adaptability, and various modular robots have been developed. Most studies have focused on the expandability of capabilities or the integration of modules, whereas only a few studies have investigated autonomous decentralized control, in which each module harmonizes its own movements for overall functionality. We developed an underwater modular robot that synchronizes its paddle strokes; the robot is based on the motif of Gonium, a multicellular alga. We built a reduced system model of modules to represent the state of an oscillator by using a phase with attractive interactions with others. Because the model is similar to the Kuramoto model, we applied analysis methods. Real robotic modules were built, and experiments were conducted using a colony of the modules. The experimental results confirmed that the colony exhibited stroke synchronization ability by compensating for individual differences. The stroke synchronization is expected to stabilize the movements of robot colonies and improve their overall propulsion.

Keywords: Multi-robot systems, swarm robots, and collaborative robots
Abstract: Robotics has been widely used in our daily life in recent years. However, programming and management for heterogeneous robot swarms are still very challenging. In this paper, we present an Actor-based programming framework for simultaneous management of multiple heterogeneous robot swarms. Our framework encapsulates robotic behavior as the ‘Actor’ by using the OODA (observation, orientation, decision, action) bus-plugin mechanism, and takes ‘Actor' as the basic control unit. Based on the proposed ‘Actor', we further designed a domain-specific language (DSL) to decouple the development into task development and algorithm development. Besides, we implement a novel communication middleware and the backbone of a multi-layer Actor model to support heterogeneous robot swarms. We implement our proposed framework in C++ and evaluate it on heterogeneous swarm platforms. The experimental results show that our framework can support the development and management of heterogeneous robot swarms.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Artificial intelligence in robotics
Abstract: Swarm robots and collaborative robots leverage cloud computing to access distributed computing, resource sharing, and collaborative support, enhancing task execution efficiency, boosting system performance, and promoting communication among robots. Task and resource allocation in cloud computing immediately impact cloud robotics systems' resource utilization and operational costs. Several metaheuristic algorithms have been suggested to improve the scheduling in cloud computing and enhance task execution efficiency. The Whale Optimization Algorithm (WOA) is a recently developed population-based intelligent optimization approach, offering a more streamlined and user-friendly implementation compared to other algorithms. It also has a loose requirement on the objective function conditions. In this paper, we employ WOA for optimizing multi-objective cloud task scheduling and propose an improved whale algorithm based on adaptive neighborhood perturbation search(ANPS-WOA) to enhance further the ability to search for optimal solutions. We detail the concrete approach for implementing ANPS-WOA and showcase through simulation experiments that it attains superior convergence speed, precision, and enhanced performance in system resource utilization.

Keywords: Space robots, aerial robots, and underwater robots, Autonomous mobile robots and manipulators, Multi-sensor data fusion and sensor networks
Abstract: Abstract—The inspection of autonomous Unmanned Aerial Vehicles (UAVs) in large cylindrical enclosed spaces has difficulties in positioning and processing massive global map data in real-time required for planning. Because such environments have the characteristics of darkness, GPS signal rejection, approximate cylinder, and large-scale environmental spaces. Therefore, we built a UAV using multi-line lidar and simultaneous localization and mapping (SLAM) in this environment to adapt to the former characteristics. In the face of the feature degradation in the height direction caused by the cylindrical-like features of the environment, the constraint information of the height sensor is added in lidar SLAM. And a map storage method called selective ordered indexed map (SOIM) is proposed for storing and real-time retrieval of huge amount of map data in large spaces required by path planning during autonomous inspection tasks. We conducted inspection experiment in a desulfurization tower of a thermal power plant to verify the function of designed system. The results prove the feasibility of our system and presented that the proposed SOIM saved 98.86% of storage space compared to the point cloud map in this environment, and the voxel search speed was 22.53% higher than that of octree search.

Keywords: Multi-robot systems, swarm robots, and collaborative robots, Space robots, aerial robots, and underwater robots, New theory and technology in robotics and biomimetics
Abstract: An unmanned aerial vehicle (UAV) close formation controller based on extended kalman filter (EKF) has been designed in this paper. The Newton-Raphson iterative method is designed for extremum location estimation, namely extremum seeking control (ESC). To facilitate extremum position estimation, the performance function is approximated using first and second-order Taylor expansions. The EKF is still utilized to estimate the gradient and Hessian matrix of the parameterized performance function. Compared with traditional estimation strategy, this allows a faster convergence and smooth command reference for close formation controller. The pigeon-inspired optimization (PIO) with pattern search (PS) is proposed to optimize these free parameters of the EKF. The simulation results show that the proposed PIO with pattern search has strong optimization performance. The EKF optimized by the algorithm enables the controller to converge well to the optimal relative position.

Keywords: Space robots, aerial robots, and underwater robots, Artificial intelligence in robotics, New theory and technology in robotics and biomimetics
Abstract: We propose an effective stall recovery learning approach based on a soft actor-critic algorithm with smooth reward functions. Stalling is extremely dangerous for aircraft and unmanned aerial vehicles (UAVs) because altitude decreases can result in fatal accidents. Stall recovery policies perform appropriate control sequences to save aircrafts from such lethal situations. Learning stall recovery policies using reinforcement learning methods is desirable because such policies can be learned automatically. However, stall recovery training is challenging since the interplay between an aircraft and its environment is very complicated. In this work, the proposed stall recovery learning approach yields better performance than other methods. We successfully apply smooth reward functions to the learning process because reward functions are critical for the convergence of policy learning. We achieve good performance by applying reward scaling to the soft actorcritic algorithm with automatic entropy learning. Experimental results demonstrate that stalls can be successfully recovered using the learned policies. The comparison results show that our method provides better performance than previous algorithms.

Keywords: Robotic vision and image processing, Space robots, aerial robots, and underwater robots, Artificial intelligence in robotics
Abstract: This paper introduces a UAV-based wheat rust detection system employing deep learning techniques. To address the limitations of traditional wheat rust disease detection methods, such as time and labor-intensive processes, lack of real-time monitoring, and excessive pesticide usage, In this paper, the YOLOv8 algorithm, representing the forefront of object detection methods is chosen for UAV wheat rust detection. We train the model on a prepared dataset. Subsequently, we evaluated 5 models, including YOLOv5, and YOLOv8, at two different scales (n, s) based on average precision (mAP@.5), precision, recall, and FPS. The experimental outcomes establish the superiority of the enhanced FasterNet-YOLOv8 model over YOLOv8 and YOLOv5 across all metrics, encompassing detection accuracy and processing speed. These findings affirm the viability of our suggested model within the UAV wheat rust monitoring system and underscore the efficacy of the model enhancements.

Keywords: Robotic vision and image processing, Multi-sensor data fusion and sensor networks, Autonomous mobile robots and manipulators
Abstract: The ever-advancing field of visual odometry has always been plagued with limited robustness. This has led to many fusion schemes being utilised in conjunction with a second modality to enhance the robustness of the visual odometry pipeline. This paper produces an extension to this body of work by being the first practical application of the stereo thermal visual-inertial navigation system based on the extended kalman filter in the literature. This paper also examines the nature of the drift inherent in the system by means of contrast to the stereo-visual and multispectral versions of the solution. All on a novel dataset fashioned to enable the study of the effects of varying illumination and heating conditions on the systems. The construction of the dataset also enabled the employment of various props to alter the feature sets present in the scene. This is used to test the effects of the varying feature sets in isolation and in addition to the various illumination and heating changes introduced to the environment.

Keywords: Space robots, aerial robots, and underwater robots, New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: This paper focuses on the autonomous navigation problem of the fixed-wing aerial robot, of which the inertial measurement unit (IMU) has time-varying error characteristic parameters. A robust variational Bayesian adaptive Kalman filter (RVBAKF) is proposed to tackle the noise outliers in integrated navigation brought by sensor uncertainties. The proposed method formulate the joint probability distribution of the system state vector and the measurement noise covariance matrix. A fading factor matrix with strong tracking quality is introduced to enhance the prediction on the prior distribution of the process noise covariance. Then a weighted sliding window mechanism has been constructed to obtain the posterior distribution of the measurement noise outliers. Therefore, the proposed approach is impressive in approximating both the process and measurement noise. The RVBAKF algorithm is implemented in an integrated navigation system which is composed of the strapdown inertial navigation system (SINS) and the global navigation satellite system (GNSS). The integration framework based on RVBAKF is conducted on two typical verification scenarios, which is proven to be exceptional in coping with the time-varying process and measurement noise by comparing the average root mean square error in misalignment angle, velocity and position with the strong tracking filter and the variational Bayesian filter.

Keywords: Autonomous mobile robots and manipulators
Abstract: Unmanned Robot Sailboats (URS) exhibit immense potential for diverse marine applications, yet their pursuit of stability often comes at the cost of sacrificing excessive speed. In response, this paper introduces a novel solution—an adaptive multihull sailboat design to reach a trade-off between stability and efficiency. Central to this paradigm-shifting concept is the convergence of a mechanical design to lift side hulls and precision control algorithms, which results in a mutually beneficial exchange. The sailboat adeptly navigates trade-offs by willingly surrendering minor speed losses for consistently stable performance. Through multiple experiments with the adjustable device in different environments and parallel comparison experiments with existing sailboats, we fully demonstrated the advantages of this boat. This paper invites exploration into a new perspective of marine robotics to achieve harmonious stability and efficiency in maritime endeavors.

Keywords: Human-robot interaction, Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics
Abstract: This paper aims to develop a simulation and visualization platform for sailing robots that enables realtime simulation, based on Unreal Engine 5(UE5), Matlab, and MySQL database. This platform provides researchers who are unfamiliar with UE5 with a convenient 3D sailing robot testing platform by creating a bridge of data communication between UE and other testing platforms using the MySQL database. During the simulation process, the sailboat dynamic model and control strategies are implemented in other platforms like Matlab, Python, or C++, which has better ability in programming and mathematic calculations. UE has a strong physics engine to do water environment simulation. A 3D sea environment for sailboats is constructed in UE. Another platform like Matlab computed sailboat state based on the sailboat model and control strategy, then transfer data to UE through MySQL database. The simulation in UE provides valuable 6 Dof of data in total including sailboat 3 dimensional position and rotation data. With real-time simulation and visualization, this platform can be utilized for testing autonomous sailboat navigation or control methods and assisting remote sea experiments.

Keywords: Applications of robotics and biomimetics, Robotics in intelligent manufacturing
Abstract: Floating ROVs have extensive applications on oceanographic research, environmental monitoring, and search and rescue operations. In this work, two levels of floating ROV stabilization is realized. First, the intrinsic stability is induced by the design of center of mass and moment of inertia based on the stability analysis, such that the ROV can always keep its upwards standing configuration. Besides, an underactuated control method for this novel untethered tumbler floating (TF) ROV is introduced for stabilization, which effectively enhances the stabilization efficiency based on the attitude controller. The model considering rigid body dynamics based on Euler equations, and nonlinear controllability with one, two, or three torques are analyzed serving for the moment of inertia design and underactuated controller. Simulation experiments are conducted to validate the efficacy of our approach, and the results demonstrate its effectiveness for stabilizing the TF-ROV.

Keywords: Autonomous mobile robots and manipulators
Abstract: This paper studies the stabilization control of underactuated Unmanned Surface Vehicles (USV). Firstly, a three-degree-of-freedom (3-DOF) of underactuated USV in complex sea conditions is established. On this basis, a backstepping sliding mode stability controller based on disturbance observer (BSMC-NDC) for USV is designed. The backstepping sliding mode control strategy is used to achieve the stabilization control effect, and the hyperbolic tangent continuous sliding mode is used to reduce the controller jitter. Aiming at the complicated ocean disturbance in the process of USV stabilization, a 3-DOF disturbance observer based on exponential convergence is designed. The effectiveness of the control system is fully verfied by comparing the simulation results of similar controllers. Specifically, simulation results show that the proposed controller can achieve the USV stabilization control and solve the jitter problem in the sliding mode control process.

Keywords: Autonomous mobile robots and manipulators, Artificial intelligence in robotics, Multi-sensor data fusion and sensor networks
Abstract: The trajectory tracking control task is of vital importance to the USV control system, especially when it is under-actuated. Focusing on the problem that the USV system model identification is difficult and the lateral motion control is usually affected by the environmental disturbance, we propose a novel USV trajectory tracking method that combines the control law with a data-driven state prediction algorithm. Due to the fact that the yaw of USV is usually calculated by a simplified dynamic model according to other state variables, accurate trajectory tracking control becomes quite difficult. Therefore, we replace the yaw calculation with a learning-based method based on the multi-sensor data captured from the USV in real- time. Our method takes the environmental factors into account while calculating the yaw variable. Simulation experiments are conducted and the comparison of 5 experimental data in the straight line and turning conditions shows that the average lateral displacement MSE error is reduced by 70.86% and 78.4% respectively compared with the model calculation method. The average tracking accuracy RMSE decreased by 43.47% and 42.31%, indicating the effectiveness of the method.

Keywords: Space robots, aerial robots, and underwater robots
Abstract: This paper designs a solar-powered autonomous underwater vehicle (AUV) with an innovative design for its recharging unit, giving it both recharging and gliding function. During the dive, the foldable charging unit is deployed and has the same gliding motion function as other solar powered underwater gliders. When it reaches the working depth, the mechanism of foldable charging unit retracts, which can shrink the area of the charging unit. This reduces the resistance of the charging plate during the underwater movement of the AUV and the force of the ocean current on the charging plate, thus saving energy and improving the stability of underwater operation. Compared with the underwater glider, the AUV in this paper is equipped with two side thrusters to control the AUV work depth, and has good depth control performance to ensure that the AUV can reach the working depth quickly and accurately. This paper discusses the structural design of the AUV, then briefly describes the dynamics modeling, and finally uses fuzzy PID control method for depth control, and conducts simulation experiments in Simulink to verify the superiority of the control method.

Keywords: Multi-sensor data fusion and sensor networks, New theory and technology in robotics and biomimetics, Applications of robotics and biomimetics
Abstract: This article discusses the challenges faced by sensor fusion systems owing to inconsistent and interrupted sensor data and proposes an adaptive polynomial predictive filter-based sensor fusion approach to address this problem. This approach incorporates an adaptive polynomial filter based on polynomial extrapolation that can effectively capture the temporal dynamics of the sensor data and provide a robust estimate of the underlying signal. This paper provides a detailed analysis of the proposed approach, including its theoretical foundation, implementation, and performance evaluation using simulated sensor data. The simulation results demonstrate that the adaptive polynomial sensor fusion approach outperforms several state-of-the-art sensor fusion techniques, particularly in scenarios in which the sensor data are inconsistent or interrupted. The article concludes that the proposed approach offers an effective solution to the problem of inconsistent and interrupted sensor data in sensor fusion, with potential applications in various fields including robotics, autonomous vehicles, and healthcare.

Keywords: New theory and technology in robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Autonomous mobile robots and manipulators
Abstract: This paper presents a novel 3D reactive navigation method for unmanned aerial vehicles (UAVs) in uneven terrain with unknown moving obstacles. The proposed method effectively reduces computational burden while ensuring safe and efficient UAV operations in complex environments. Through extensive simulations, the research demonstrates the efficacy of this approach. The findings lay a solid foundation for practical UAV applications in challenging 3D environments

Keywords: New theory and technology in robotics and biomimetics, Space robots, aerial robots, and underwater robots
Abstract: To achieve Mars rotorcraft attitude control with improved control accuracy and immunity, an improved African vulture optimization algorithm (IAVOA) is proposed and applied to the attitude controller parameter adjustment. Considering the coupling effect of control inputs and the uncertainty of the Mars environment, the attitude is controlled using an active disturbance rejection controller after decoupling the model using state feedback. In order to obtain the optimal parameters of the controller, the optimization algorithm is improved by using tent mapping initialization, opposite learning strategy, memory search strategy, weight time-varying mechanism, Cauchy Mutation, and attraction mechanism for the problem that the AVOA (African vulture optimization algorithm algorithm) is prone to fall into local optimal solutions and slow convergence of the algorithm. The simulation is verified with the ITAE (Integral of Time-weighted Absolute Error) index as the optimization target and compared with PSO (Particle Swarm Optimization), CSA (Crow Search Algorithm), and AVOA. The results demonstrate that the IAVOA converges faster, finds better parameters, and improves the control accuracy and immunity to disturbance.

Keywords: New theory and technology in robotics and biomimetics
Abstract: Unmanned Aerial Vehicles (UAV) are widely used in military and civilian applications as an emerging control technology. Due to the stringent coordination requirements and constraints, multi-UAV path planning in three-dimensional complex environments is more challenging. To solve this problem, A collaborative method for planning multiple UAVs paths in complex environments is proposed. The SMA algorithm is improved in terms of the initial position of the population, the feedback factor and the individual position update method, which improves the convergence speed and convergence accuracy as well as the robustness of the algorithm. Simulation results in complex cases show that the proposed algorithm can obtain effective paths in multi-UAV collaborative path planning, and the overall performance is improved by 9.5% compared to the SMA algorithm.

Keywords: Space robots, aerial robots, and underwater robots, Robotic vision and image processing
Abstract: In this paper, we propose and design an underwater delta robot for fast seafood grasping. First, the hardware structure of the robot is described in detail. After that, a visual servo control method for fast catching of this underwater delta robot is proposed. The method is able to generate real-time radial trajectories so as to realize the catching based on the swaying of the robot body as well as the movement of the object to be caught. In the actual grasping test, the moving platform and the slave arm can occlude to the target resulting in the loss of target position information. Therefore, we propose a position prediction method to predict the position of the grasped object when occlusion occurs, thus improving the success rate of grasping and ensuring a smooth robot trajectory. Finally, several land and underwater experiments were conducted with good results, which verified the feasibility of the robot structure and algorithm.

Keywords: Space robots, aerial robots, and underwater robots, Multi-sensor data fusion and sensor networks, Smart sensors and actuators
Abstract: With the increasing prevalence of drones in various industries, the navigation and tracking of unmanned aerial vehicles (UAVs) in challenging environments, particularly GNSS-denied areas, have become crucial concerns. To address this need, we present a novel multi-LiDAR dataset specifically designed for UAV tracking. Our dataset includes data from a spinning LiDAR, two solid-state LiDARs with different Field of View (FoV) and scan patterns, and an RGB-D camera. This diverse sensor suite allows for research on new challenges in the field, including limited FoV adaptability and multi-modality data processing.

The dataset facilitates the evaluation of existing algorithms and the development of new ones, paving the way for advances in UAV tracking techniques. Notably, we provide data in both indoor and outdoor environments. We also consider variable UAV sizes, from micro-aerial vehicles to more standard commercial UAV platforms. The outdoor trajectories are selected with close proximity to buildings, targeting research in UAV detection in urban areas, e.g., within counter-UAV systems or docking for UAV logistics.

In addition to the dataset, we provide a baseline comparison with recent LiDAR-based UAV tracking algorithms, benchmarking the performance with different sensors, UAVs, and algorithms. Importantly, our dataset shows that current methods have shortcomings and are unable to track UAVs consistently across different scenarios.

Keywords: Space robots, aerial robots, and underwater robots, New theory and technology in robotics and biomimetics
Abstract: 在自治的传统路径规划中 水下航行器（AUV），算法的时间 并且更多地考虑了规划路径的长度。能源 消耗极大地影响AUV的续航能力，是 很少 考虑。因此，一种基于路径规划算法 本研究提出了能源消耗。首先， AUV在导航过程中的能耗函数是 既定。然后，根据能耗 功能 A*算法的成本函数是自适应的 改进。 因此，一种具有最优能量的改进型A*算法 建议消费。之后，算法，A* 算法、蚁群算法和改进蚁群 算法 在光栅地图下模拟。仿真结果 表明与其他三种算法相比， AUV在规划的路径上航行的能耗 算法减少6.0%∼22.9%。

Keywords: Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics
Abstract: Regular defect inspection of the power line using X-ray is essential for the maintenance of the power line. Usually, conducting such an inspection with a wheeled robot requires dragging the robot from the ground and carefully placing it upon the power line, which is laborious and unsafe. To improve inspection efficiency, the newly developed unmanned aerial vehicle (UAV) provides a promising alternative. However, the positioning error from Global Navigation Satellite System (GNSS) brings the small-scaled drifting movement of the UAV and X-ray camera system, which leads to imaging blur. To cope with this issue, we design a flexible towed aerial robot system to alleviate the instability of the X-ray camera system. Specifically, the UAV and X-ray camera carrier are flexibly connected by a cluster of ropes, reducing the physical impact from the small-scaled drifting movement of the UAV. The permitting position error tolerance between the UAV and the carrier is analyzed. In addition, a guide wheel frame is designed on the carrier to facilitate the carrier's smooth rolling along the power line. Furthermore, aiming to adapt to the different types of power lines, we design a lightweight motor-driven system to adjust the camera angles and the imaging plate position. Multi-view cameras are also designed to assist the pilot to control the UAV carrying the X-ray camera system landing on the power line. To verify the performance of the developed aerial robot system, we conduct real-world experiments with double bundle conductors and four bundle conductors. The results show that the developed system can efficiently complete inspection. The X-ray camera could obtain a stable imaging condition under the small drifting movement of the flight.

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: This paper reports the design innovation and implementation of a novel wheel-leg transformable robot named SWhegPro3. The robot uses a three-impeller design to significantly improve the overall stair-climbing performance. Unlike other impeller-based designs that use gear sets to unfold the impellers, this robot uses self-locking electric push rods for mode switching, thereby greatly improving its load capacity. We also analyze the stair climbing operation and derive a model to calculate the optimal configuration of the wheel-leg module and the corresponding displacement of the push rod, in order to better adapt to different sized stairs. We conduct both simulation and field experiments to validate our proposed design and control strategy.

Keywords: Medical robotics, biomedical and rehabilitation engineering, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Applications of robotics and biomimetics
Abstract: Colonoscopy is a widely used method for diagnosing bowel cancers, requiring coordinated efforts from medical professionals due to its invasive nature. In contrast, miniature capsule robots, measuring just a few centimeters, offer a less invasive diagnostic option by integrating multiple functions. However, current capsule robots face challenges with complex designs, limited mobility, size, and cost, which hinder their adoption and reduce gastrointestinal disorder screening rates. To address these issues, we introduce the vibration-driven capsule robot (VCR) with magnetic protruding bristle structures. Four simple components (shell structure, electronic components, permanent magnet, and magnetic soft bristles) are designed to improve the robot's mobility in dry and wet environments. The magnetic bristle structures are inspired by flexible organisms and enhance the robot's adaptability. The robot is propelled by centrifugal force from a vibration motor, allowing efficient movement along the intestinal tract while protecting the patient's lining. The fabrication process of the magnetic soft bristles has been refined to optimize material ratios and enhance the robot's performance. The vibration motor's output is analyzed using load cells to ensure efficiency. For precise path planning, an external permanent magnet (EPM) is applied to guide the capsule's direction. Experiments confirm the VCR's versatility, performing fluid linear and curved motions in different environments. These findings underscore the feasibility of the VCR as a potent tool for non-invasive, timely screening of gastrointestinal disorders.

Keywords: New theory and technology in robotics and biomimetics, Space robots, aerial robots, and underwater robots, Applications of robotics and biomimetics
Abstract: A lunar regolith in-situ analysis payload deploying a mass spectrometer is proposed to exactly detect the water ice of the South Pole of the moon for China future lunar exploration missions. In order to realize the goal, an airtight heating furnace for volatile extraction from lunar regolith is desired. In this paper, sealing methods for lunar regolith volatile in heating furnace to future proposed in-situ sampling and analysis of lunar samples were investigated based on power requirements, furnace closing accuracy requirements, dust tolerance and operation in the Lunar environment. Methods implemented include a balanced metal hollow ring seal, a double channel ferrule tab ring radial seal and a spring energized seal. Comparing the performance of sealing methods under the above conditions, a spring energized seal is selected as the heating furnace seal. The materials and structure of the seal are designed based on the rigidity requirements, leakage rate requirements and working environment requirements. In order to test the sealing performance of seals after repeated sealing in high and low temperature environments, a seal leak detection system was designed. The test results indicate that the seal can meet the sealing performance requirements after repeated sealing at various temperatures.

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: Traditional quadruped robots typically have rigid and fixed trunk, which limit their mobility and flexibility. This paper introduces a novel quadruped robot with a reconfigurable trunk and presents an adaptive gait design on slope with maximum stability margin. The analysis of stability margin in multi-leg robots is significant for achieving smooth walking performance. By adjusting the trunk folding angle using the waist motor, the proposed gait enables the robot to achieve optimal stability margin during walking on slope. The strategy was thoroughly investigated and experimentally validated on slopes. The experimental results demonstrate that employing the designed gait, the robot’s stability margin during walking on a slope of 15°is enhanced by 24.7% compared to the conventional walk gait. Furthermore, the robot can utilize the waist motor to transform between different biomimetic configurations, enhancing environmental adaptability and expanding its range of applications.

Keywords: Robotic vision and image processing, Artificial intelligence in robotics, New theory and technology in robotics and biomimetics
Abstract: In recent years, deep learning based feature fusion has drawn significant attention in the field of information integration due to its robust representational and generative capabilities. However, existing methods struggle to effectively preserve essential information. To this end, this paper proposes a gate-based fusion module for object detection to integrate the information from distinct feature layers of convolutional neural networks. The gate structure of the fusion module adaptively selects features from neighboring layers, storing valuable information in memory units and passing it to the subsequent layer. This approach facilitates the fusion of high-level semantic and low-level detailed features. Experimental validation is conducted on the public Pascal VOC dataset. Experiments results demonstrate that the addition of the gate-based fusion module to the detection task leads to an average accuracy increment of up to 5%.

Keywords: Human-robot interaction, Artificial intelligence in robotics, Applications of robotics and biomimetics
Abstract: The emergence of large language models (LLMs) has paved the way for advancing robotics capabilities, especially in intricate tasks that demand nuanced comprehension and precision. In this context, this paper introduces a novel interactive camera robot that harnesses LLMs to enhance human-robot interaction and optimize robot control. Specifically, an innovative technique that leverages the language understanding capabilities of LLMs to plan camera movement trajectories and waypoints is presented. In this study, the geometric relationships among the objects under capture are employed to plan the control strategy. Accordingly, this approach not only empowers sophisticated camera parameter manipulation and color adjustments but also fosters a natural and efficient human-robot interaction. Lots of experiments on real robots are conducted to evaluate the effectiveness of the proposed method under various scenarios. The results reveal robust performance across crucial measures, affirming the substantial potential of LLMs in elevating camera robot control and interaction experience. Videos of our experiments are available at https://youtu.be/zP-sTZHvXe4.

Keywords: Robotics in intelligent manufacturing, Applications of robotics and biomimetics, Autonomous mobile robots and manipulators
Abstract: Progress monitoring is crucial for the construction industry. It directly impacts the construction period and affects project cost and quality. A lot of previous research focused on in-process progress monitoring. They compared the point cloud of the main structure with the BIM model to check the gap between as-built and as-planned status. This article proposes a BIM-based after-process progress monitoring system based on a visual method. It aims at indoor components after the main structure built. The robot can navigate to the installation sites automatically on a BIM-generated map and check the status of components. We propose a novel and lightweight method for object recognition and estimation of the three-dimensional position of objects relative to BIM coordinate. It is realized by RGB-D camera and deep learning. The semantic and geometric information of BIM is utilized to filter out false recognition of deep learning. The final detection results are compared with BIM to generate a progress-monitoring report. Finally, we build a real robot and check the installation status of four components in a corridor. And all the components are checked successfully.

Keywords: Robotic vision and image processing, Multi-sensor data fusion and sensor networks, Artificial intelligence in robotics
Abstract: The application of robots in social life, equipped with sensors and actuators and embedded with AI, assists people in all aspects. However, the first perspective of the robot horizon is heavily constrained, which weakens its performance. A joint tracking system is designed and built to deal with this, by integrating a surveillance system with the robot visual, providing a third perspective. This system takes one horizontal view and two top views from various directions as inputs and matches a person among the frames and in time sequence. In order to deal with the identity match with a huge visual feature gap, a special dataset is collected, simultaneously labeling identities from a mobile robot perspective and multiple indoor static surveillance monitors. The experiment shows that such match is a task worth exploring that can be better handled by training on our dataset than existing open source Re-identification (Re-id) datasets. Moreover, in the real scenario, this system improves the performance on issues like in and out of the robot's field of vision and heavy occlusion by people or objects.

Keywords: Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: Dynamic locomotion of a quadruped robot, ensuring stability and agility, is crucial for practical use. Our research group has devised an elastic leg mechanism inspired by the muscle-tendon complex of the gastrocnemius and the achilles tendon in the lower limbs of animals. While we have accomplished dynamic locomotion with our quadruped robot, integrating motion control that considers its behavior remains crucial to harness the advantages of this biological normative mechanism. Gait control, implemented through central pattern generator (CPG) that provides feedback on environmental interactions detected using toe force sensors, holds the potential to actively incorporate the influences of leg mechanisms that inherently respond to movement in locomotion control. In this paper, we proposed a leg elastic force feedback CPG model aimed at effectively utilizing the bioinspired leg mechanism to quadrupedal locomotion by realizing coordinated movements among the limbs of a quadruped robot, instead of using force sensors at toes. This coordination is based on the state of the elastic leg spring and the phase of the toe trajectory. The proposed model enabled the expansion of the stable running region and facilitated faster movements during locomotion by utilizing elastic force feedback.

Keywords: Autonomous mobile robots and manipulators, Multi-sensor data fusion and sensor networks, Robotic vision and image processing
Abstract: Autonomous movement of legged robots in outdoor scenes requires accurate state estimation and environment navigation maps. Different from wheeled and tracked mobile robots, legged robots can adjust parameters such as gait type, step height, gait frequency, torso height, foot friction, and other coefficients to better adapt to the geometric characteristics and physical properties of verstile terrains. To this end, in this paper, we propose a framework for autonomous navigation of legged robots, including the construction of geometric-property RGB consistent semantic map (RCS) and the autonomous navigation method under the RCS map. The method for building RCS map formulates lidar-inertial odometry based on a factor graph. Through end-to-end semantic segmentation of images, terrain information is projected and saved while mapping. For achieving pose and point cloud consistency, local and global lidar boudle adjustment (BA) is introduced to the optimization process. We demonstrated the effectiveness of RCS through outdoor navigation experiments on a quadruped robot. The experiment shows that our terrain detection method can achieve an average detection time of 30 ms on an embedded board.

Keywords: Robotics in intelligent manufacturing, Multi-sensor data fusion and sensor networks, Applications of robotics and biomimetics
Abstract: This paper introduces a new High-speed Alignment Software adopting Servo-based approach. It builds on the External Guided Motion (EGM) interface of ABB robots. A novel framework for data fusion and servo is proposed with the emphasis on its improvement to traditional approaches and adaptability to sensors with low frame rate (around 10Hz). Auto spatial- and time-synchronization functions help user to deploy the software with less effort and expertise. This software can be applied in both static alignment application and target tracking (dynamic alignment) cases. Experiment results show that accuracy for static alignment can be improved by 50% and time consumption reduced by 70%. And error for dynamic alignment is still at the same level when sensor frame rate is reduced from 200Hz to 15Hz. The proposed solution can be applied in various applications including 3D gluing, fly-picking and other tasks requiring ultra-high precision.

Keywords: Autonomous mobile robots and manipulators
Abstract: In this paper, a novel bilateral teleoperation control is designed specifically for manipulation in contact environment. Firstly, serval interaction conditions (e.g., free motion and rigid interaction) are defined and unified into one equation as the designed unified interactive model, which can also be utilized as a control objective. Based on this model, a remote hybrid motion/force controller is designed for remote robot, which achieves the good position and force tracking, and ensures the precise and safe contact force that helps reduce the workload of operator. Subsequently, the contact force is reconstructed in the local side by estimating and transmitting the non-power environment coefficient from remote side, which provides the operator with accurate force feedback. Since the coefficient replaces the direct transmission of force signal in the communication channel, the traditional power-cycle problem is essentially avoided, which guarantees the stability of teleoperation system under communication time delays. Comparative experiments are implemented to verify the effectiveness of proposed method for teleoperated manipulation.

Keywords: Autonomous mobile robots and manipulators
Abstract: Robot grasping has always been a hot topic in the research field, but there are still a lot of problems to be solved. In this paper, we propose a general grasping method for the unknown objects in cluttered environment from a single-view. Our method is divided into three steps: sampling points selection, grasping pose sampling and grasping pose evaluation. Firstly, we generate a pixel-wise probability image based on the trained sampling affordance network from RGB images, and map the points with high probability to the 3D point cloud combined with the depth map. Then, we obtaine the candidate grasps by sampling on the proposal points, and finally we evaluate the candidate grasps through the trained grasp evaluation network. The network can capture the geometric structure represented by the point cloud information. The proposed method directly uses the original sensor data and can increases the sampling success rate.

Keywords: Smart sensors and actuators
Abstract: We propose a novel grasp detection method for arbitrary shape objects. This method is applicable to multijoint manipulators with a two-finger hand and an RGB-D hand-eye camera. When a robot tries to grasp an unknown object with a camera as the only sensor, it must choose a position to grasp without mass distribution information. Moreover, even if the robot can find some candidates of locations for grasping, some of them will be false due to noise. The proposed method chooses an appropriate candidate by variational Bayesian clustering. In the clustering, the candidates are classified based on their closeness. Then the method chooses one existing at the center of the largest cluster since it may be uninfluenced by noise. This set of procedures does not require training unlike learning methods. The proposed method is evaluated with an actual robot. Though the method is a heuristic, it can choose suitable grasping locations from piles of objects.

Keywords: Autonomous mobile robots and manipulators, Applications of robotics and biomimetics
Abstract: In this paper, we study the mobile manipulation problem with a quadruped-arm system in cluttered enviornment. To address this problem systematically, we propose a real-time collision-free motion planning and control framework that distributes the collision avoidance task into multiple layers. In details, we decompose the overall legged mobile manipulator system into a quadruped subsystem and an arm subsystem, plan the collision-free trajectories for the base and the end-effector while respecting their inherent kinematic constraint. Then with the guidance of the base and end-effector trajectories, we can find feasible velocity commands for the quadruped and the arm respectively by solving a quadratic programming problem with the collision-free constraint to guarantee safety of the whole body system. After getting the velocity commands for the quadruped, we further synthesize a customized model predicitve controller to track the given reference while incorporating the effect of the moving arm. The framework is validated by the navigation and manipulation tasks in simulation with a real legged-mobile manipulator system. In the experiments, the average computation time for our motion planner is 76 ms and the planned trajectory is smoother and faster than the existing motion planners. Simultaneously, the collision avoidance requirement can be obeyed during the whole process.

Keywords: New theory and technology in robotics and biomimetics, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Autonomous mobile robots and manipulators
Abstract: Ants use their antennae as tactile probes to sense their environment, then execute grasps using their mouth-parts on a variety of novel objects. In this work, we define kinematic and motion constraints to create a simplified ant model, which is used to evaluate different hypothetical approaches to tactile probing and grasp synthesis and identify plausible descriptors for real ant behaviour. Using this model, we compare several methods for obtaining tactile data, and for model-free grasp synthesis from this gathered data. We evaluate each combination according to several measures of grasp quality, and the relative cost to execute each behaviour on the robot. The best results are obtained using a PCA-based grasp synthesis method and this is most robust across various shapes if previous contact points influence the selection of future locations to probe.

Keywords: Autonomous mobile robots and manipulators
Abstract: In this paper, robotic regrasping is considered with the goal of achieving dexterous manipulation. The strategy using quick wrist snap is based on human regrasping, and involves rotating the grasped object due to the inertial force while maintaining the contact state. The posture of the grasped object is measured by high-speed vision in real time and is used for posture control. Experimental results are shown where regrasping of a stick-type object in the direction contrary to gravity can be achieved by a high-speed robot hand.

Keywords: Soft robotics and liquid-metal robotics, Smart sensors and actuators, Applications of robotics and biomimetics
Abstract: By using a mechanism that dynamically changes morphology, it would be possible to always perform the optimal morphological processing, e.g. hardware filtering in a task. In this study, we proposed a soft mechanism in which the morphology of the finger pad changes, in this case with variable stiffness, and investigated the relationship between the stiffness of the finger pad and classification accuracy in an object classification task using dynamic touch. The proposed mechanism was implemented on the finger pad of a two-finger gripper and shaken while grasping a PET bottle filled with different contents. As a result, we confirmed that the time series information obtained from the torque sensor at the wrist changed depending on the stiffness of the finger pad, and that the content classification results using the time series information changed.

Keywords: Artificial intelligence in robotics, Autonomous mobile robots and manipulators
Abstract: Abstract—Redundant manipulators are often required to not only track target trajectories but also actively avoid obstacles in practical applications. A real-time trajectory planning method is proposed to solve the issue of redundant manipulator obstacle avoidance and tracking. Firstly, the method takes the joint angle limit formed by the physical mechanism of the manipulator itself and the obstacle limit in the workspace as constraint conditions. Additionally, the optimization objective is the error between the current trajectory and the desired trajectory, and quadratic programming technology is used to describe it. Furthermore, the method employs a recurrent neural network (RNN) for real-time optimization on the quadratic programming model. Finally, the effectiveness of the proposed method is verified by experiments on Franka manipulator. Experimental results demonstrate that the algorithm is highly efficient for manipulating trajectory planning, suitable for solving the problems of obstacle avoidance and joint angle limits while tracking trajectories.

Keywords: Robotics in intelligent manufacturing
Abstract: The automation of garment manufacturing has been intensified due to labor shortages and amplified production needs. The handling and manipulation of fabric alone consume nearly 80% of the total production line time. Hence, there is a significant research focus on developing specialized grippers for fabric handling. This paper discusses the design of an under- actuated robotic hand tailored to efficiently handle garments and perform post-production tasks, such as integration into

ironing or screen-printing machines. To evaluate its effective- ness, experiments simulating the task of placing garments onto a rectangular plate, akin to a screen-printing machine, were conducted.

Keywords: Soft robotics and liquid-metal robotics, Applications of robotics and biomimetics, New theory and technology in robotics and biomimetics
Abstract: Robotic grippers are used in a wide range of industrial applications for the manipulation of objects. To simplify the assembly process of robotic grippers, compliant grippers with a monolithic structure are introduced in the gripper design. In this paper, we propose a unified surface-model-based framework in MATLAB to achieve the parametric design of 3D-printable compliant grippers. A compliant parallel-grasping mechanism is utilized in this work to achieve parallel closing motions of the gripper jaws, while the shape of the jaws is adapted to the target object to realize task-specific grasping. As a proof of concept, two grippers designed by using our framework are used to grasp objects with different shapes. Results show that the printed gripper prototypes can firmly grasp the objects, which successfully verified the feasibility of our design framework.

Keywords: Robotics in intelligent manufacturing, Applications of robotics and biomimetics, Multi-sensor data fusion and sensor networks
Abstract: The manipulator systems with vision perception play an import role in intelligent manufacturing, helping to achieve industrial automation production. However, the electronics manufacturing industry, especially in scenarios such as consumer electronics manufacturing, has a strong demand for flexible manufacturing. The traditional calibration methods for vision perception require manual operation, customized calibration objects and their world-frame coordinates, which cannot meet the requirements of flexible production. Therefore, a novel agile vision calibration, perception and control algorithm is designed to meet flexible manufacturing requirements in the paper. The proposed algorithm receives vision pixel raw information as unique input, derives calibration matrix and manipulator control target pose automatically. The algorithm is adapted for both eye-in-hand and eye-on-hand vision sensor, and the theoretical analysis shows that the calibration algorithm has no manual operation errors. Experimental verification shows that the designed calibration algorithm in this article is simpler to cooperate, has fewer sources of error, easier to control, and has higher calibration accuracy than traditional calibration algorithms, while meeting the requirements of flexible manufacturing.

Keywords: Robotic vision and image processing, Multi-sensor data fusion and sensor networks, Artificial intelligence in robotics
Abstract: As a representative bunch-type fruit,the collision free and undamaged harvesting of grapes is of great significance.To obtain accurate 3D spatial semantic information,this paper proposes a method for multi-feature enhanced semantic segmentation model based on Mask R-CNN and PointNet++.Firstly,a depth camera is used to obtain RGBD images.The RGB images are then inputted into the Mask RCNN network for fast detection of grape bunches.The color and depth information are fused and transformed into point cloud data,followed by the estimation of normal vectors.Finally,the nine-dimensional point cloud,which include spatial location,color information,and normal vectors,are inputted into the improved PointNet++ network to achieve semantic segmentation of grape bunches,peduncles,and leaves.This process obtains the extraction of spatial semantic information from the surrounding area of the bunches.The experimental results show that by incorporating normal vector and color features,the overall accuracy of point cloud segmentation increases to 93.7%, with a mean accuracy of 81.8%.This represents a significant improvement of 12.1% and 13.5% compared to using only positional features. The results demonstrate that the model method presented in this paper can effectively provide precise 3D semantic information to the robot while ensuring both speed and accuracy. This lays the groundwork for subsequent collision-free and damage-free picking.

Keywords: Robotics in intelligent manufacturing, Robotic vision and image processing, Artificial intelligence in robotics
Abstract: With the rapid development of artificial intelligence and computer vision, numerous technologies have been introduced to automate manufacturing in the industrial domain. Typical metal workpieces in the industry often have highly reflective surfaces, come in various sizes, and are positioned irregularly. The motor rotor presented in this paper is one such representative workpiece. Traditional grasping methods for workpiece loading and unloading are pre-programmed and often struggle to cope with complex and disordered situations. In this paper, we introduce a structured light (SL) sensor as the visual guide for the triaxial robot. Furthermore, we propose a high-precision hand-eye calibration method for the non-orthogonal coordinate system of the triaxial robot. Additionally, a motor rotor center localization method based on U-Net image segmentation is proposed. By combining the high-precision hand-eye calibration and localization, we can accurately and automatically locate and grasp the rotor. We have conducted sufficient experiments to verify the effectiveness and accuracy of our system.

Keywords: Artificial intelligence in robotics, Robotic vision and image processing, Robotics in intelligent manufacturing
Abstract: Since the current grasping success rate of robots is low when performing grasping tasks in complex environments, in order to improve this problem, this paper proposes a robot grasping detection network SA-U2GNet combining U2-Net and Shuffle Attention networks. The network can not only achieve information communication between different sub-features through the attention mechanism, but also capture more contextual information from RGB-D images through the two-level nested U-shaped structure. Training and testing were performed on the Cornell and Jacquard grasp datasets, the accuracy rates reached 97.9% and 94.7% respectively, and the time required to process RGB-D images was 30ms. Compared with other methods, this method improves the accuracy and time efficiency, and the experiment verifies the feasibility and effectiveness of this method.

Keywords: Autonomous mobile robots and manipulators
Abstract: This paper describes a motion generation method for a robot performing door-opening tasks. Unlike prior studies that divided the door-opening task into subproblems, we propose a single computation approach to obtain robot motions. To achieve this, we construct a robot model utilizing differential algebraic equations (DAE) derived from the differential kinematics of mobile manipulators. Additionally, we present the formulation of a nonlinear optimization problem, including the definition of constraints for robot’s via points, collision avoidance, and others. We then present our experimental findings conducted in both simulated and real environments, involving various types of hinged doors. We confirmed that the proposed method can generate suitable motions for door opening, both for the robot itself and for others.

Keywords: Robotic vision and image processing, Artificial intelligence in robotics, Smart sensors and actuators
Abstract: Visual servoing has been gaining popularity in various real-world vision-centric robotic applications. Autonomous robotic grasping often deals with unseen and unstructured environments, and in this task, Visual Servoing has been able to generate improved end-effector control by providing visual feedback. However, existing Servoing-aided grasping methods tend to fail at the task of grasping in dynamic environments i.e - moving objects. In this paper, we introduce DynGraspVS, a novel Visual Servoing-aided Grasping approach that models the motion of moving objects in its interaction matrix. Leveraging a single-step rollout strategy, our approach achieves a remarkable increase in success rate, while converging faster and achieving a smoother trajectory, while maintaining precise alignments in six degrees of freedom (6 DoF). By integrating the velocity information into the interaction matrix, our method is able to successfully complete the challenging task of robotic grasping in the case of dynamic objects, while outperforming existing deep Model Predictive Control (MPC) based methods in the PyBullet simulation environment. We test it with a range of objects in the YCB dataset with varying range of shapes, sizes, and material properties. We show the effectiveness of our approach by reporting against various evaluation metrics such as photometric error, success rate, time taken, and trajectory length.

Keywords: Robotic vision and image processing, Human-robot interaction, Robotics in intelligent manufacturing
Abstract: This paper proposes a robotic automated layup framework consisting of three components: human-robot layup skill transfer, vision-based quality detection, and adaptive impedance controller. The framework aims to address challenges and limitations faced by manual carbon fiber layup, such as high labor intensity, poor quality consistency, and limited scalability. In the proposed framework, firstly, on a horizontal plane, humans demonstrate the layup behavior by dragging the robot in gravity compensation mode. During the demonstration, the Cartesian position, orientation, and interaction forces of the robot's end-effector are recorded synchronously. Secondly, utilizing the Mask R-CNN instance segmentation model, defects such as bubbles and wrinkles on the carbon fiber surface are detected. Depth feature extraction and bounding boxes are employed to determine the position and orientation of the defects. Lastly, based on hand-eye calibration and coordinate system transformation, the robotic adaptive impedance controller repeats the layup behavior demonstrated at the defect pose points for automated defect repair. The feasibility and effectiveness of the proposed framework were validated in real-world scenarios, demonstrating its suitability for automated layup tasks.

Keywords: Soft robotics and liquid-metal robotics
Abstract: Although various grasp strategies have been proposed in soft gripper area, form-enclosing grasping mode always plays a marginal role because of the challenges to calculate the deformations of soft materials. In this study, we propose a easy-to-make kirigami gripper and a geometric optimization method to obtain the fingers boundary shape that achieves a complete form-enclosing grasping after bending. We used a FEM model to represent the finger shape and calculated the finger boundary, and then used a iterative method to find the best finger shape that generates the form-enclosing grasping result. The optimization method is based on the minimum energy rule, and has the advantages of being computationally efficient and being universal for optimizing the boundary of gripper with form-enclosing grasping mode. The effectiveness of the proposed method was validated by ANSYS simulation and practical experimental results. Moreover, an objective functions and multiple geometric constraints were derived to quantitatively evaluate the gripper designs.

Keywords: Human-robot interaction, Bio-inspired robots, e.g., climbing, creeping, and walking robots, Medical robotics, biomedical and rehabilitation engineering
Abstract: Hand gestures are a natural way of communication and integrating them into robots could allow for more efficient human-robot collaboration. In recent years, researchers and roboticists have attempted to replicate human hand motor control using the concept of synergies. In this paper, we present a new approach to obtaining kinematic synergies from hand gestures using a single RGB camera. We capture real-time hand gestures using the MediaPipe framework and convert them to joint angular velocities. We then use dimensionality reduction to obtain kinematic synergies from the joint angular velocities. These synergies can be used to reconstruct new hand gestures. We translate these reconstructed hand movement patterns into a humanoid robot, Mitra. Our results show that it is possible to control most of the joints of the robot for performing hand gestures using only a few synergies. This is more efficient than other contemporary methods. Furthermore, robots and prosthetics that use synergy models could enable near-natural human-robot collaboration.

Keywords: Smart sensors and actuators, Applications of robotics and biomimetics
Abstract: Tactile sensors are crucial technology for autonomous operations of robots. In this paper, we propose a 3×3 array fingertip tactile sensor based on the Velostat pressure-sensitive material. To address the issue of abrupt signal variation in the micro-force response stage, a mechanical structure applying preloading force to the sensing units is designed. Besides, in order to enhance the pressure transmission to the sensing units, silicone contacts are manufactured to obtain significant signal changes. We built the specialized experimental platform to test the performance of the designed sensor. The experimental results show that the sensor can respond to the compression and release of the silicone contact. We also use the sensor to contact the object and successfully obtaining the contour feature information of the object. This study will be helpful for designing the better fingertip tactile sensor.

Keywords: Smart sensors and actuators, Multi-sensor data fusion and sensor networks
Abstract: Vision-based tactile sensors (VBTS) leverages visual modality to present high-resolution tactile information. The vision-based sensing mechanism has good adaptability to robot manipulation because of available information capture and data processing. However, the simulation of VBTS is a challenging problem because it deals with computer graphics and elastoplastic deformation. In this paper, we propose a simulation approach for VBTS using a rendering approach based on path tracing. The rendering method can be integrated with existing VBTS and robotic simulators for the simulation of robotic systems. In addition, this method can control the rendering efficiency and quality by controlling the number of voxels in the fitting deformation area to meet the requirements of efficient robot training. The experimental results verify that our method can provide high-quality images at low rendering efficiency or reduce image quality to improve rendering efficiency. Our work has the potential to advance the sim-to-real research on VBTS.

Keywords: Applications of robotics and biomimetics, Space robots, aerial robots, and underwater robots, Soft robotics and liquid-metal robotics
Abstract: In this study, a flexible grasping terminal manipulator that uses the pneumatic vacuum principle and granular material as its base is given. Pneumatic device and granular material are wrapped in elastic film by the end-effector. Whatever complex shape structure the target object has, the flexible end can change shape and wrap around it to entirely grasp the object, it is easy to operate and effective. Create a model of a multi-finger paw with flexible ends that may be used on other things with form structures in addition to flat objects. The flexible end effector has improved application prospects and practical relevance thanks to the optimized design, which additionally offers high gripping speed, success rate, and stability.

Keywords: Robotics in intelligent manufacturing, Robotic vision and image processing
Abstract: The majority of current bin picking systems, designed for industrial parts, cannot be directly oriented to the downstream task after grasping. This research presents a grasping framework that addresses this challenge by incorporating pose estimation of parts in cluttered bin environments and the targeted design of robot end-effector grippers. This approach ensures that the pose of the part on the gripper is known and fixed, enabling successful assembly tasks in various scenarios. To train an object pose estimation network, we propose a system for generating a dataset of industrial parts using model rendering within a physics engine. We analyze the geometric features of the parts, and further design a gripper, to achieve the grasping strategy. Results demonstrate that for a single known industrial part, the minimum grasping success rate is 91.4% in simulated robot experiments, and the assembly success rates in different scenarios based on this framework exceed 80%. Our framework offers valuable guidance for the deployment of robotic grasping.

Keywords: Human-robot interaction, Autonomous mobile robots and manipulators
Abstract: In hazardous explosive disposal scenarios, the autonomous grasping system of quadruped robots demonstrated limited task-level decision-making capabilities and exhibited minimal human-robot interaction (HRI). This paper proposes a task-level intelligent HRI approach, aiming to assist the quadruped robots making multi-object grasping decisions. Following the principle of natural HRI, we design a control terminal equipped with a touch screen. The control terminal receives the video stream from the robot's primary field of view via a video transmitter. We develop a graphical user interface (GUI) that decodes and displays the image, allowing the operator to interact naturally by touch and select a specific point in the image as the search center point. We employ the shortest Euclidean distance principle, enabling the target recognition module to search around the selected center point. Subsequently, the autonomous grasping system determines the grasping target and initiates the tasks of path planning and grasping detection. Real-world experiments involving multi-target autonomous grasping are conducted to validate the effectiveness of the proposed HRI method and address the challenge of decision making in scenarios with multiple targets. Furthermore, this method can be readily extended to other types of mobile robots.

Keywords: Space robots, aerial robots, and underwater robots, Autonomous mobile robots and manipulators
Abstract: In the context of long-term experiments within a space experimental module, the assistance of space robotic arms becomes crucial for astronauts. Considering the significant contributions of redundant robotic arms in various fields, a compact, hyper-redundant, and highly flexible robotic arm is designed for these space missions. To address the inverse kinematics problem for this specific structure, a C-IK algorithm is proposed, which combines the Non-Dominated Sorted Genetic Algorithm-II (NSGA-II) and Sequential Quadratic Programming (SQP) methods for multi-objective optimization. By modifying the algorithm's structure, C-IK retains global search capability, significantly improves runtime efficiency, and meets accuracy and continuity requirements. Experiments involving C-IK, NSGA-II, and an improved version of NSGA-II were conducted on 7-Degrees-of-Freedom (DOF), 8-DOF, and 9-DOF robotic arms to evaluate their performance. The results demonstrate that C-IK outperforms NSGA-II in terms of both algorithm runtime and accuracy, especially for the 9-DOF robotic arm, with the runtime reduced from over 37 seconds to just over 2 seconds and the accuracy increasing from less than 10% to higher than 95%. The findings highlight the significance of the proposed C-IK algorithm in achieving efficient and accurate inverse kinematics solutions for hyper-redundant robotic arms in space missions.

Keywords: Smart sensors and actuators
Abstract: Soft tactile sensors are a family of versatile flexible sensors aiming to replicate the sense of touch using skin-like artificial systems, with the primary function of contact localization. Such devices can rely on a multitude of physical principles and structural designs. Most implementations exploit contact-induced changes in electrical or optical resistivity, achieving good performance, but strongly limiting the range of usable materials. On the other hand, solutions relying only on structural design and delocalized sensors have yet to be shown able to achieve the same performance. In this paper, we propose a barometric soft tactile sensor that utilizes a grid of sealed cavities and delocalized barometric sensors to perform contact localization independently from the contact depth. Moreover, since the device's function is not material-specific, we also investigate the role of compliance by testing multiple increasingly softer materials. For each material, we characterize the cavities' sensing field and then collect the data sets to train and test a multi-layer perceptron regressor for contact localization, both with a constant and variable contact depth. The results show that we are able to lower the contact localization mean error down to 2 mm.

Keywords: Autonomous mobile robots and manipulators, Robotic vision and image processing, Artificial intelligence in robotics
Abstract: In this study, we develop an automatic control system to perform the reach-to-grasp movement of a 7-DOF (Degrees of Freedom) robotic arm that has the same DOFs as a human arm, and an end-effector with the same shape as a human hand. The 6-DOF pose of the object to be grasped is estimated in real time only from RGB images using a neural network based object pose estimation model. Based on this information, motion planning is performed to automatically control the reach-to-grasp movement of the robotic arm. In the evaluation experiment, the 7-DOF robotic arm performs reach-to-grasp movements for a household object in different poses using the developed control system. The results show that the control system developed in this study can automatically control the reach-to-grasp movement to an object in a certain arbitrary pose.

Keywords: Applications of robotics and biomimetics, Robotic vision and image processing, Robotics in intelligent manufacturing
Abstract: Robotic manipulation systems are urgently necessary to replace the labor generated by a avalanche of repetitive test tube operations in pharmaceutical, chemical,and biological fields. However, the currently existing test-tube manipulation robot [1] have poor generality and inefficiency in object detection and grasping reasoning. To this end, we propose a robotic manipulation system for test tubes arrangement harnessing geometric insights to enhance generality and efficiency. For object detection, instead of template matching in [1], we propose a corner positioning method combining vision and depth to detect positions of the rack and its holes; For grasp reasoning, poses of the gripper are adaptively adjusted by the estimated attitudes of test tubes, instead of brute force search in the working space as in [1]. Experiments manifest our robotic manipulation system to be more highly generalized and efficient in contrast to [1], achieving success rates for randomly-posed and vertical-posed test tubes up to 96.6 % and 100 % respectively, and the execution efficiency is improved by 38.4%.

Keywords: Artificial intelligence in robotics, Robotic vision and image processing, Applications of robotics and biomimetics
Abstract: Active object detection (AOD) allows robot to approach target objects smoothly, but previous studies have assumed that the robot can always see the object from the start. This study introduces a new AOD model that can learn to spatial explore from any position to complete AOD task using a deep recurrent Q-learning network (DRQN). The model combines historical with current state information to accurately predict the action that should be performed at the current moment. This study also introduces a new reward function and termination condition for AOD task to ensure task completion rates and efficiency. This approach has been validated on an active visual dataset (AVD), demonstrating that service robot can actively complete AOD task from random positions through spatial exploration. Our approach has also been found to be superior to other methods.

Keywords: New theory and technology in robotics and biomimetics, Smart sensors and actuators, Bio-inspired robots, e.g., climbing, creeping, and walking robots
Abstract: The heat generated inside robots originates mainly from energy loss in actuators, onboard electronic circuits, and computation processing units. Typically, it can be dissipated by forced air (most commonly used in robotics) or fluid convection. In this study, we developed a recycled inner-generated heat system inspired by biological thermoregulation mechanism, utilizing the inner heat to the robot's thermal perceptions of a finger-pad (by analogy to human fingertip) for object haptic recognition. Among the fingertips' thermal perceptions, material recognition identifies and distinguishes touched objects, even if the color, stiffness, or roughness are similar. This recognition approach requires a heat source to induce temperature changes at the contact surface to recognize ambient (room) temperature objects. We use an actuator that generates heat in a closed-flow water circuit (by analogy to human cardiovascular system) to induce heat at contact like humans with body temperature and touch. Our thermal method is assessed through experimental simulations of robotic water circulation and a pump system (by analogy to human heart) with the developed finger-pad. The proposed strategy enables it to completely classify three kinds of material covering the same material in 0.7 sec touch.