Keywords: Control Theory and Technology, Automation Systems, Mechatronics Systems
Abstract: In the field of engineering, the stability and performance evaluations of control systems are crucially important. Micro-controllers and digital processors play a vital role in the design and implementation of control systems due to their high-speed and precise information processing abilities and their capability to perform complex calculations in real time. Therefore, to comprehend the performance and behavior of control systems implemented by digital processors, analyzing them in the discrete domain is an essential factor. Compared to the continuous-time model, the discrete-time model provides a more accurate representation of the behavior of these systems. To analyze and design digital DOB-based control systems, it is essential to consider an appropriate discretization method. This paper investigates the performance of DOB-based speed control systems in the discrete-time domain, providing a more comprehensive understanding of the dynamic response of digital motion control systems compared to continuous-time analysis methods. The theoretical and experimental analyses show that the selection measurements or estimator, along with the discretization method for DOB synthesis, significantly affects the stability and performance of digital motion controllers.

Keywords: Human-Robot Cooperation/Collaboration, Human-Robot/System Interaction, Human Interface
Abstract: Prosthetic hands utilizing electromyography have been developed so far. Recently, highly sophisticated prosthetic hands with multiple degrees of freedom have emerged. However, the number of movements estimable from myopotential signals is limited, approximately a dozen or so. Hence, the development of interfaces capable of controlling multiple degrees of freedom is anticipated. Our research group proposes a control system integrating vision sensors with myopotential signals. However, methods based on deep learning pose challenges due to their high computational demands. This paper presents a practical approach developed using a vision sensor, a microcontroller, and a lightweight myopotential sensor.

Keywords: Autonomous Vehicle Navigation, Machine Learning, Motion and Path Planning
Abstract: Using machine learning to process image data from cameras provides a more intelligent approach to autonomous navigation. However, navigation in snow environments remains a challenging problem since it is difficult for traditional machine learning approaches to handle noises caused by falling snowflakes. In this study, we propose a a novel approach called adversarial reservoir computing (ARC) for the mobile robot to tolerate interference from falling snowflakes and achieve autonomous navigation safely in snowy environments.

Keywords: Robotic hands and grasping, Soft Robotics, Mechanism Design
Abstract: This paper presents the rationale behind the development of a stiffness-controlled soft suction gripper based on layer jamming effect, highlighting its motivation and distinguishing features in comparison to prior research. It offers insights into experimental findings and discusses both the limitations encountered and potential solutions.

Keywords: Micro/Nano Systems, Medical Systems, Hardware Platform
Abstract: This study focuses on improving magnetic manipulation systems for microrobot actuation. We introduce a novel approach for globally optimizing the electromagnet configuration by integrating finite element methods (FEM) in the Bayesian optimization loop. The actuation performance is improved by 20% compared to previous magnetic dipole model-based optimization method.

Keywords: Robotic hands and grasping, Biologically-Inspired Robotic Systems, Soft Robotics
Abstract: This study aims to investigate the role of the fingernail-like structure in soft artificial fingers integrated with a vision-based tactile sensor (VBTS). The evaluation test is conducted on two features: 1) stability in object grasping task and 2) tactile sensing capability.

Keywords: Multi-Robot Systems, Robotic hands and grasping, Machine Learning
Abstract: We present a novel framework, Demo Augmented Transformer Reinforcement Learning (DA-TRL), to learn multi-agent long-horizon (LH) tasks with a dual-arm robot. DA-TRL uses a transformer policy and human demos to learn the state-action space for multi-agent tasks and a human-like object manipulation skill. DA-TRL outperforms conventional methods, achieving a 65% success rate improvement in complex LH tasks.

Keywords: Rehabilitation Systems
Abstract: More than 40 % of postpartum women are experiencing urinary incontinence, which is a significant challenge for postpartum women. In this paper, we developed a system to evaluate the stiffness of the pelvic floor muscles. We verified its operation in order to assess functional impairment of the pelvic floor muscles, which is one of the causes of urinary

Keywords: Modeling and Simulating Humans, Human Factors and Human-in-the-Loop, Human-Robot/System Interaction
Abstract: In home environments, fragile things and narrow spaces inside furniture are used by humans and have to be used by robots. To deploy robots in houses, human motion features and home furniture configurations were analyzed. We focus on the careful manipulation motions of the humans handling a fragile cup in a space-limited storage box. Based on the measured data using motion capture system and six-axis load cell, we calculate twisting angles of the torso, external and internal angles of the shoulder, and flexion and extension angles of the elbow for evaluation. To evaluate the careful manipulation motion, we configured three conditions (25%, 50%, and 100%) of placeable areas (PA). On the narrow 25% PA, twisting angles of the torso, external and internal angles of the shoulder, and flexion and extension angles of the elbow were increased. During approaching the object on the bottom of the storage box, we confirmed that the target orientation of the right hand was achieved by twisting angles of the torso, external and internal angles of the shoulder, and the target depth position of the right hand was achieved by flexion and extension angles of the elbow.

Keywords: Mechatronics Systems, Mechanism Design, Motion and Path Planning
Abstract: In this paper, a climbing robotic wheelchair mechanism with a multi-functional balancing leg can be used as both the counterweight for the two-wheel balancing mode and the leg for climbing stairs. These two legs make a gate motion when the robot climbs stairs. These two legs help to balance the robot when it is operating in balancing mode on flat ground, as well. We verified the validity of the proposed robot mechanism and tested the climbing motion using the prototype hardware.

Keywords: Vision Systems, Sensor Fusion, Automation Systems
Abstract: In this paper, we introduce a vision-based recognition method of the stairs for various climbing robotic systems. In order to climb the stairs of the robotic system, information on the stairs such as the width and the height of the stair step is needed. In the proposed method, we used facet information of the stairs obtained from the stereo vision and depth cameras to recognize the stairs. In order to verify the feasibility of the proposed method, we tested the proposed algorithm with a small size of the stair model.

Keywords: Soft Robotics, Mechanism Design
Abstract: This short paper presents our attempt to construct a dynamic model of a deformable propeller when colliding with an external object. The mathematical model describes the deformation of soft parts of the propeller and the brushless direct current motor (BLDC) motor. Besides, the aerodynamic aspects of the wing and impact force are also considered in this model. To evaluate the proposed model, we plan to execute the comparison study between simulation results, including the position and velocity of the propeller and experimental results of the propeller’s behavior upon impact in reality.

Keywords: Human Interface, Virtual Reality and Interfaces
Abstract: Emotion estimation based on facial electro- myography (fEMG) during Virtual Reality (VR) experience is useful. A wearable fEMG device which can be used with commercial VR headsets enables to apply to various applications, however, there are few such devices. In this study, we confirmed that both a developed wearable device and a VR headset can be attached to a participant and the fEMG could be measured during the VR experience.

Keywords: Control Theory and Technology, Machine Learning, Automation Systems
Abstract: This paper proposes a new approach to realize a force sensor-like reaction force observer based on deep learning. The reaction force observer is designed by a LSTM deep neural network model. To reproduce the reaction force similar to the one measured by force sensor, the deep neural network is trained by force information from force sensor, position, and reference current of the motor. Then, only position and motor current information is employed for force estimation when implementing the proposed force observer. The effectiveness of the proposed method is verified by numerical simulation results.

Keywords: Rehabilitation Systems
Abstract: Rehabilitation for stroke patient is difficult due to various impairments affect motor function, motivation, and sense of agency (SoA). SoA is important because it is related to motivation and performance. However, it is difficult to evaluate SoA nonbinary and easily. The aim of this study is to investigate the nonbinary measurement of this sense by measuring the left prefrontal cortex and right inferior parietal lobe activity using functional Near-Infrared Spectroscopy.

Keywords: Autonomous Vehicle Navigation, Vision Systems, Machine Learning
Abstract: In ground-view object change detection, the recently emerging mapless navigation has great potential to navigate a robot to objects distantly detected (e.g., books, cups, clothes) and acquire high-resolution object images, to identify their change states (no-change/appear/disappear). However, naively performing full journeys for every distant object requires huge sense/plan/action costs, proportional to the number of objects and the robot-to-object distance. To address this issue, we explore a new map-based active vision problem in this work: ``Which journey should the robot select next?" However, the feasibility of the active vision framework remains unclear; Since distant objects are only uncertainly recognized, it is unclear whether they can provide sufficient cues for action planning. This work presents an efficient simulator for feasibility testing, to accelerate the early-stage R&D cycles (e.g., prototyping, training, testing, and evaluation). The proposed simulator is designed to identify the degree of difficulty that a robot vision system (sensors/recognizers/planners/actuators) would face when applied to a given environment (workspace/objects). Notably, it requires only one real-world journey experience per distant object to function, making it suitable for an efficient R&D cycle. Another contribution of this work is to present a new lightweight planner inspired by the traditional multi-armed bandit problem. Specifically, we build a lightweight map-based planner on top of the mapless planner, which constitutes a hierarchical action planner. We verified the effectiveness of the proposed framework using a semantically non-trivial scenario ``sofa as bookshelf".

Keywords: Autonomous Vehicle Navigation, Logistics Systems, Automation Systems
Abstract: This research deals with navigation, mainly trajectory planning and tracking for mobile robot in warehouses, considering safety and travel strategies. Conventional navigation systems generally integrate the measurement results of on-board sensors into a map, and the navigation system commands the robot to avoid or decelerate in response to obstacles. However, due to the effects of map resolution and measurement errors, the distance to obstacles in real space and on the map may diverge, which is an issue from a functional safety point of view. Therefore, recent industrial robots are being considered to be equipped with safety LiDAR to ensure higher safety functions. The safety LiDAR dynamically changes its detection range according to the robot's velocity measured by an on-board encoder. Since, the robot operates independently of navigation, it is difficult for conventional navigation systems to achieve optimal travel control, taking in account of changes in the detection range. In addition, in actual warehouse operations, operators must change the robot's travel strategy from time to time depending on the on-site conditions, such as work efficiency, energy consumption, and load stability. Therefore, this study proposes a navigation system that can flexibly respond to travel strategies, considering the safety LiDAR. To be Specific, we focused on Model Predictive Control, which can explicitly handle a wide variety of constraints, and designed objective functions for multiple travel strategies considering the predictions of the safety LiDAR behavior. To reduce the computational load, the system was designed so that safety LiDAR behavior and travel strategies are considered only in the trajectory planning performed by the server PC, while the on-board PC only focuses on following the target trajectory. This allows the robot to only follow the planned trajectory and to adapt to various travel strategies in real time, taking into account the behavior of safety LiDAR.

Keywords: Autonomous Vehicle Navigation, Machine Learning, Vision Systems
Abstract: This work introduces a comprehensive deep reinforcement learning (DRL) framework for autonomous driving (AD) in high-fidelity traffic environments. The fusion of the robust open-source platforms CARLA and RLlib establishes a unified and customizable framework for DRL applications in AD. It provides a transparent and objective testing of the models, and further enables users to customize components such as the neural networks, reward functions, and input data. Moreover, it allows structured, reproducible assessments through the CARLA leaderboard, providing critical metrics, including lane violations, collision incidents, and route length. Through rigorous, objective evaluations, we ascertain the efficacy of our DRL models in independently navigating an ego vehicle under the demanding conditions simulated by CARLA. The cornerstone of this research is the comparative analysis conducted between standard end-to-end DRL and our proposed decoupled DRL strategy, further enriched with advanced techniques such as imitation learning and data fusion. Our findings show the strengths of the decoupled DRL approach—wherein the perception task is separated from the decision task—in driving the ego vehicle, thereby achieving state-of-the-art performance. We also highlight the pivotal role of advanced strategies, particularly imitation learning and data fusion, in bolstering performance and addressing intricate tasks. Moreover, our framework provides an easy-to-use and standardized approach to explore vision-based and decoupled DRL under Partially Observable Markov Decision Process (POMDP) conditions with spatiotemporal inputs in complex AD scenarios.

The framework is fully open source and can be found on GitHub: https://github.com/hallilibas-if/ddhRL.

Keywords: Autonomous Vehicle Navigation, Sensor Fusion
Abstract: This study proposes an approach for robust localization with low computational cost in autonomous vehicles. We introduce EKF localization that combines NDT and DR. The uncertainty estimation considering NDT localization errors caused by environmental features is challenging in EKF localization. Although an approach using multiple matching processes from offset initial positions is effective in tackling this problem, it remained pre-computation due to its high computational cost. To achieve real-time performance, we arrange offset initial positions only in the direction where the NDT localization error is likely to expand by using the Hessian matrix of the NDT score function. Moreover, the number of initial positions increases only when the NDT localization error is expanded. To further reduce the computational cost, this study also proposes hybrid correspondence NDT (HC-NDT). HC-NDT reduces the number of unnecessary correspondences of neighbor searches while maintaining the robustness and accuracy of conventional NDT. Our EKF localization with HC-NDT and our real-time uncertainty estimation suppresses errors within 0.3 m in challenging environments for NDT, such as tunnels. The execution time of our uncertainty estimation with HC-NDT is reduced by up to 50 % compared to using conventional NDT.

Keywords: Autonomous Vehicle Navigation, Machine Learning, Motion and Path Planning
Abstract: Navigation systems for mobile robots using deep reinforcement learning have the potential to achieve autonomous movement without relying on precise maps. Therefore, we are conducting a study on autonomous navigation of mobile robots using monocular camera images as input through deep reinforcement learning. Typically, trained models based on deep reinforcement learning are trained in simulation environments. However, there are gaps between simulation and real-world environments, making it challenging to apply trained models to actual robots. This challenge is particularly pronounced when using images as input. Semantic segmentation is employed to address this issue, as it can simplify complex RGB images into segmented images, thereby reducing the gaps between environments. In this study, we acquire a model for mobile robots to reach their destinations and acquire a semantic segmentation model. These models are applied to a ROS-based autonomous navigation system. Real-world experiments confirm the successful application of the learned models in actual environments.

Keywords: Autonomous Vehicle Navigation, Vision Systems, Sensor Fusion
Abstract: In this system paper, we present our approach to the DJI Robomaster AI Challenge, 2022 where our team ERA-IITK was ranked 3rd worldwide among 83 teams. We primarily focus on system hardware design and algorithmic components to achieve full autonomy. We first detail the hardware infrastructure, including intricate details on sensors and communication setup. Then we comprehensively describe our algorithmic pipeline and a state machine based on them to realize an autonomous system. Our major novelty is the high-speed and accurate visual perception module for target detection, and high-speed self-localization to maximize the likelihood of a target being shot. Our research endeavor in this paper can serve as a foundation for more complex systems and holds immense potential for various real- world applications.

Keywords: Human-Robot Cooperation/Collaboration, Virtual Reality and Interfaces
Abstract: This paper presents a novel approach that combines mixed reality technology with variable impedance control to enhance the collaboration between humans and robots. The HoloLens 2 mixed reality headset is used to collect task-related information, such as goal positions and environment mapping. This spatial information is then used to calculate a distance field, which provides precise distance measurements from any position to the nearest object in the environment. The proposed method focuses on adjusting impedance parameters, including damping and stiffness, based on the generated distance field. By incorporating this environmental information, the proposed method improves the operational performance of human-robot collaborative tasks. Moreover, the use of HoloLens 2 mixed reality capabilities allows the proposed method to be effortlessly applied to new environments or goals. Experimental validations were conducted using a 7-degree-of-freedom (7DOF) robot, demonstrating the effectiveness of the proposed method in enhancing human-robot collaboration.

Keywords: Human-Robot Cooperation/Collaboration, Integration Platform, Multi-Robot Systems
Abstract: Digital Twins (DTs) have been used as enablers for applications in optimization, monitoring, and other analyses along the lifecycle of their physical counterparts. In the robotics domain, DTs have been used as high-level components to connect robotic control with extended services. However, DTs do not necessarily include a structured taxonomy for the so-called operations, although it has been approached in the robotics domain, such as the skill-based engineering concept. Additionally, representing particular cases, such as cooperative robotics, is not an easy task, especially due to the complexity involved in system composition and coupled behavior.

In this paper, we propose an extension to a modeling approach for composed DTs in cooperative systems that adopts the skill-based engineering concept. The extension ties to an ontological model that represents DTs in four components, namely, attributes, operations, behaviors, and relationships. The operation component is then extended by including the skill-based concept and inheriting the ontological properties of the base modeling approach.

This contribution is evaluated using a case study with two cooperative robotic arms. The evaluation consists of two experiments, which are compared in two different settings, a manual fashion implementation and an implementation following this approach. The results show i) an improvement in terms of implementation effort and ii) enhanced reusability when switching to different tasks via the skill-based-enabled DTs.

Keywords: Human-Robot Cooperation/Collaboration, Human-Robot/System Interaction, Automation Systems
Abstract: In order to realize human-robot cooperative work, it is necessary to have dynamic interaction for unexpected human's motion that are difficult to define in advance. In this research, we propose a robotic system that supports operator work without interfering with human work by reacting and avoiding operator motion through real-time motion generation and activation re-planning. This system is included 3D recognition of the robot's surroundings and real-time motion generation technology. We implemented in the setup of assembly work assuming a high-mix, low-volume production site, and was demonstrated in a support task that robot distribute the parts for operator's assembly work. Experimental results show that the robot can conduct the cooperative work without interfering with operator work by performing appropriate avoidance actions against operator motion.

Keywords: Human-Robot Cooperation/Collaboration, Human-Robot/System Interaction, Software Platform
Abstract: Natural language is an effective tool for communication, as information can be expressed in different ways and at different levels of complexity. Verbal commands, utilized for instructing robot tasks, can therefor replace traditional robot programming techniques, and provide a more expressive means to assign actions and enable collaboration. However, the challenge of utilizing speech for robot programming is how actions and targets can be grounded to physical entities in the world. In addition, to be time-efficient, a balance needs to be found between fine- and course-grained commands and natural language phrases. In this work we provide a framework for instructing tasks to robots by verbal commands. The framework includes functionalities for single commands to actions and targets, as well as longer-term sequences of actions, thereby providing a hierarchical structure to the robot tasks. Experimental evaluation demonstrates the functionalities of the framework by human collaboration with a robot in different tasks, with different levels of complexity. The tools are provided open-source at https://petim44.github.io/voice-jogger/

Keywords: Human-Robot Cooperation/Collaboration, Human-Robot/System Interaction, Human Interface
Abstract: In recent years, the emergence of cooperative robots has been facilitated by technological advances, including the development of safety devices. Although these robots are considerably more flexible than conventional industrial robots, when confronted with a new task, the human operator is often left to address related problems unaided. The aim of this research is to develop a smart collaborative robot, where a rapid and accurate robot and a flexible human operator work synergistically to achieve a common objective. In the proposed system, humans and robots work together to tackle the same combinatorial optimization problem; the robot actively engages in problem-solving, while the human has the ability to override the robot's actions if disagreements arise, and to execute alternative actions. To facilitate this collaboration, this paper presents the development of a manipulator that enables two separate control centers, the human brain and the robot's computational system, to cooperate in solving complex, cognitive-motor tasks.

Keywords: Human-Robot Cooperation/Collaboration, Human-Robot/System Interaction, Human Interface
Abstract: In this study, we are verifying the use of the developed cooperative robot to safely perform lifting tasks with humans. Currently, robots are expected to be utilized to address the challenges posed by labor shortage and in improving productivity. However, the functions of robots are overwhelmingly limited compared to those of workers, and they are restricted to applications where their strengths can be utilized. Therefore, this study focuses on lifting work by two workers and attempts to develop a system to perform lifting work with one worker and one robot. One of the challenges of this work is how to reflect the unconscious actions of humans in the cooperative robot. Therefore, by deducing the velocity displacement of lifting motions performed by humans and reflecting them in the robot, we confirmed the contribution to smooth lifting operations by humans and cooperative robots.

Keywords: Hardware Platform, Systems for Field Applications
Abstract: A FPGA-based channel estimation design is presented to combine the fine symbol time offset (FSTO) and the residual carrier frequency offset (RCFO) algorithms. The proposed architecture is highly scalable, supporting multiple antenna configurations and varying pilot symbol lengths. In this paper, a full design flow from algorithm analysis to implementation and optimization is described with FPGA demonstration. The 2x4 MIMO Uplink system has been implemented in Xilinx UltraScale VCU118 to achieve a low latency computation time of 0.392ms, less than the 0.5ms 5G transmission time interval (TTI), leaving rooms for other computations and system up-scaling.

Keywords: Integration Platform, Network Systems
Abstract: In recent years, there has been an increasing demand to adopt renewable energy resources to power grids for environmental issues and energy concerns. A virtual grid system, which is designed to utilize decentralized resources, is proposed as an approach. This system enables users to construct their power grids with virtual grid hubs (VG Hub) which interconnect via commercially available cables according to USB Power Delivery (USB PD). A prototype of VG Hub is capable of synthesizing and distributing DC power for connected devices, and the power flow control is managed by a microcontroller. Our previous study proposed a method called Minimum-Hop Power Path Routing (MHPPR). This method determines the power flow that uses the closest source in candidates which can fulfill the demand when a load connects to the network to minimize power loss due to transmission. However, this could result in unproductive power flow, such as a flow that consumes power stored in batteries prior to surplus power from generators. Therefore, we assumed that there are factors to value a source for a load other than demand and supply capacity, and inferred that the factor derives from the type of device. We defined this factor as the affinity between devices. The objective of this study is to solve the power path optimization problem while taking affinity into account. In this study, we discussed device classification and affinity between types and defined affinity coefficients. We then extended MHPPR by incorporating this affinity coefficient. The results have shown that the proposed method selected the sources following the priority defined by the affinity.

Keywords: Integration Platform, Software Platform, Control Theory and Technology
Abstract: In practical applications, only the output noise, but not the disturbance of the system, can be measured directly. In this case, the effectiveness of the designed Kalman filter is poor and even diverges because of the disturbance. In this paper, we make use of the improved Sage-Husa state disturbance statistical estimators to estimate the mean and variance of the system state disturbance in real time. To provide more robustness, the strong tracking Kalman filter algorithm is introduced to improve the variance of state prediction in time. For the velocity model of the gyro-stabilized platform with different positive and negative velocity model parameters, we verify the superiority and practicability of the algorithm proposed through comparative experiments under different system conditions in the simulation experiments. This paper designs a better Kalman filter with process disturbances and noise and provides a deep study and investigation by means of a comparative analysis of performance.

Keywords: Integration Platform, Mechatronics Systems, Systems for Field Applications
Abstract: Workers are highly impacted by work-related musculoskeletal disorders in several industry positions. To mitigate this risk of injury, a potential solution is adopting active occupational exoskeletons in the workplace. Because these wearable devices are task-dependent, the user may need access to some domains to configure the exoskeleton parameters. The User Command Interface is a wearable human-machine interface, it is used as an adaptable setup system for occupational exoskeletons. It grants secure user identification, user capture information, and exoskeleton asset configuration such as controller gain and torque limits. The occupational exoskeleton Shoulder-sideWINDER is designed to modulate proper assistive forces for upper-limb users. This paper presents an evaluation of the human-machine interface integrated into the exoskeleton Shoulder-sideWINDER. The user can quickly make easy adjustments to adapt the assistive forces according to the task it is performing. A total of 16 participants conducted standard overhead lifting experiments for the Shoulder-sideWINDER exoskeleton. Results show that the User Command Interface assessment satisfies diverse usability attributes such as complexity, learnability, ease of use, intuitiveness, performance, reliability, and satisfaction.

Keywords: Hardware Platform, Automation Systems
Abstract: In developing mobile robots for exploration on the planetary surface, it is crucial to evaluate their performance, demonstrating the harsh environment in which the robot will actually be deployed. Repeated experiments in a controlled testing environment reproducing various terrain and gravitational conditions are essential. This paper presents the development of a minimal and space-saving indoor testbed, which can simulate steep slopes, uneven terrain, and lower gravity, employing a three-dimensional target tracking mechanism (active xy and passive z) with a counterweight.

Keywords: Integration Platform, Automation Systems, Vision Systems
Abstract: Abstract— This video demonstration paper presents an integrated vision-based control framework for imitation learning and adaptive manipulation control of robot arms. It addresses the challenges of programming robot skills for industrial applications. The results illustrate that the framework enables our robotic arm system to passively observe user demonstrations and imitate the demonstrated (complex) trajectories in pick-and-place tasks. Furthermore, the robot can automatically adapt its motion on the fly to a change in the target position (placement) and avoid an obstacle.

Keywords: Mechatronics Systems, Mechanism Design, Integration Platform
Abstract: This paper proposes a prismatic joint mechanism composed of timing belts and pulleys to enhance the wheel-legged locomotion of a mobile quad-arm robot proposed by the authors. The previous mobile quad-arm robot had four arms composed of revolute joints only, and the pushing force was insufficient to walk and lift the torso simultaneously. In this study, the third revolute joints of the two rear arms are replaced with the prismatic joints. The arm with the prismatic joint mechanism achieves a maximum force of 222 N independent of other joint angles. The maximum force is enough to lift the robot while the robot walks, in which the robot's weight is almost 18 kg. We validated the prismatic joint mechanism via a simulation and experiment and conducted four tasks: standing, lateral wheel-legged locomotion, forward wheel-legged locomotion, and object grasping with the mobile quad-arm robot.

Keywords: Mechatronics Systems, System Simulation
Abstract: The increasing need for storing goods efficiently in warehouses while maintaining the same floor area raises the challenge of designing vehicles with high lift capacity and with a high flexibility at the same time. Modern Reachtrucks with lift heights of up to fifteen meters belong to this category. This causes a vulnerability to structural bending of the material which can lower workspace safety and induce damage to the goods. To cope with this problem, mast damping systems are necessary. For a proper active mast damping by the forklift's actuators a precise knowledge of the system behavior in push direction, i.e. the system's eigenfrequency, is necessary. In the industry, a common procedure is therefore a time and resource intensive measurement of the system's eigenfrequencies for different lift heights. This study provides a precise mathematical formulation of the mast system as a flexible multibody system based on easy accessible parameters which models the eigenfrequencies sufficiently. The system's equations of motion are derived by formulating the total energy of the system and making use of the Lagrangian method. In the end, measurements with different masts of different heights and lift capabilities verify the presented method.

Keywords: Mechatronics Systems, Control Theory and Technology, Sensor Networks
Abstract: The performance of mobile robot localization is highly dependent on the accuracy of motion and measurement models. In real-world situations, the actual motion of a mobile robot may differ from the control input (command) due to slippage or skidding of wheels, which causes errors in the motion model and leads to degradation of localization accuracy. This study presents a new robust localization algorithm that integrates the finite memory structure with unbiased weighted Frobenius norm (UWFN) estimation for mobile robots with uncertain control inputs. The effect of uncertain control input was minimized by UWFN estimation, and the accumulation of errors in the measurement model was suppressed by the finite memory structure. The performance of the proposed algorithm was verified through experiments using an ultra-wideband-based wireless sensor network and a mobile robot.

Keywords: Mechatronics Systems, Mechanism Design
Abstract: Synthetic fiber ropes are used in many robots because of their light weight and high tensile strength. However, the coefficient of friction of synthetic fiber ropes is low, resulting in the difficulty of rope fixation under high tension. This study focuses on the rope fixation method using a fixed pulley. From Euler's belt theory and the relationship between D/d and tensile strength, the shape of the pulley that is equal to the tensile strength is calculated as a non-circular curve. Based on this curve, we propose the rope fixation method using a non-circular pulley that enables the area of a fixed pulley to be reduced. Fixation by proposed non-circular pulley achieved 74-100% load capacity relative to the original tensile strength.

Keywords: Mechatronics Systems, Energy and Environment-Aware Automation, Hardware Platform
Abstract: In this paper, we proposed a concept of passivecooling for robot actuators which dissipates generated heat among components by thermal conduction. Graphite sheets were adopted as a substance of thermal conduction and those effectiveness were studied through experiments using a motor and the sheets. As a concrete implementation of passivecooling using graphite sheets, a robot module with passivecooling concept was developed. Through the experiments of temperature measurement, we confirmed that the rise of motor temperatures were reduced in the situation with sheets for thermal conduction.

Keywords: Entertainment Systems, Virtual Reality and Interfaces
Abstract: This study investigated the effects of two techniques to improve the sense of immersion and flying for flight simulators in immersive virtual reality environments. To test their effects, we used a drone simulator, with an operator standing on a drone and controlling it over a natural landscape by inclining his/her body toward the travelling direction. The first technique involved vibratory feedback through a foot plate that delivered the simulated vibration and sounds of the left and right propellers of the drone. The second technique involved the passive roll and pitch rotation of the foot plate when the operator actively shifted the center of his/her body mass. In a user experiment involving 12 participants, the passive inclination of the stage was found to enhance immersive feelings, while the vibratory feedback to the soles were found to enhance both feelings of flying and immersion. This study helps develop a drone and other flying simulators using immersive virtual reality settings.

Keywords: Decision Making Systems, Control Theory and Technology, Machine Learning
Abstract: Robot control systems built with flexibly combinable sub-task modules can be applied to diverse and complex tasks by reusing or recombining those modules. Previous studies have proposed multiple module designs to achieve a goal task, but they require learning or manual design of dependencies between modules, and in a number of cases, learning the order of execution of modules toward a goal, which is a challenge for a composable robot control system in diverse applications. To solve this problem, we propose goal-oriented task planning based on inter-module similarity. As transitions between modules are possible when the termination of one module is similar to the initiation of another, our method defines the state distributions at the initiation and termination of each module, and automatically estimates the dependencies between modules on the basis of their similarity. We further developed a goal-oriented task planning method by incorporating these modules into the framework of Control as Inference [1], [2]. We evaluated our method in an environment simulating a kitchen task using an arm robot and showed that it performs as well as or better than conventional methods without manual design and learning of dependencies between modules.

Keywords: Environment / Ecological Systems, Systems for Field Applications, Factory Automation
Abstract: Designing and integration of cooperative Information Systems is crucial in one of the significant endeavors of this time: decarbonization. This article addresses the integration of resilient infrastructure and secure instruments for monitoring, reporting and verification (MRV) of greenhouse gas (GHG) emissions, and lights on a scenario in the artisan coffee industry. We present a ongoing field study with three industry partners, and the goal is to design a effort sharing consortium, which employs a blockchain-based edge infrastructure to securely connect organizations and their production machinery to enable cooperative automation of MRV for GHG emissions.

Keywords: Environment Monitoring and Management, Software Platform, Sensor Networks
Abstract: Air pollution is of great concern for governmental agencies, business sectors and the public due to its adverse impact on human wellness. Effective monitoring and forecasting of air pollution are essential for stakeholders to take necessary actions for risk management. This paper presents the development of a user-friendly visualization platform that can incorporate deep learning estimation schemes in ensemble with observations and modeling data for air pollution forecast globally, and real-time monitoring of those airborne pollutants at a local scale. A pipeline is developed to directly transport the forecast data from the extended model and cloud-based real-time ambient data from low-cost sensors. The resulting dashboard offers the possibility to customize various options such as the regions of interest, timescopes of forecast, air pollutant types, and the selected models. Interactive and engaging features designed for the platform promote users' experiences to explore the rendered data from a diversity of air quality models and sources. This integrated solution aims to provide stakeholders with a holistic yet explicit perspective on the spatial-temporal disperses of air pollutants with accuracy and reliability. Such insights into the ambient conditions can contribute to raising environmental awareness, providing input to microclimating, enhancing the ability to make informed decisions, analyzing meteorological patterns, supporting policy formulation, and facilitating proactive climate responsiveness.

Keywords: Environment Monitoring and Management, Systems for Field Applications
Abstract: In this study, we evaluate efficient traveling strategies for a water surface robot to cover multiple aquatic environment observation points. In the previous study, we have developed an autonomous water surface sensing robot called BIWAKO-X and evaluated efficient position-keeping strategies for long-term textit{in-situ} aquatic environment observation. In this paper, we focus on a multi-point traveling strategy that is designed for comprehensive environmental observation across multiple locations. Assuming a scenario in which the robot travels to several observation points, the travelling strategy was constructed with observation cycles and travelling site distances. In simulation and practical field experiments, traveling strategies applying different observation cycles and travelling site distances were evaluated in terms of energy consumption. Through both simulations and hands-on field experiments, we examined traveling strategies. The results reveal that varying the observation cycle and traveling site distance affect power consumption and spatio-temporal observation resolution.

Keywords: Intelligent and Flexible Manufacturing, Integration Platform, Systems for Service/Assistive Applications
Abstract: The increasing prevalence of 3D food printing technology is shaping culinary cultures, including the creation of noodles with specific thickness and form. However, a critical challenge exists in assessing the textural characteristics or chewiness ("koshi") of these 3D-printed noodles, making it unclear whether they offer novel sensory experiences or mimic existing noodles. This study proposes a system to quantitatively evaluate and extrapolate the chewiness of 3D-printed noodles, using the "Implantable Gel Biter” which is a piezoelectric sensing mastication robot that mimics the human oral structure. These technologies work synergistically, allowing for the accurate assessment of noodle quality based on chewiness, with a focus on texture and elasticity. The developed system, underpinned by soft matter based reservoir computing, successfully identified the firmness characteristics of noodles with an accuracy rate of 81.9%, indicating its effectiveness in discerning variations in maximum stress in different noodles. This breakthrough could lead to the development of a new food texture analysis device.

Keywords: Intelligent and Flexible Manufacturing, Decision Making Systems, Mechatronics Systems
Abstract: Growing demand for new buildings and infrastructure, combined with stagnating productivity, is making new innovations and automation in the construction sector crucial. Already during the design and planning phase the feasibility of the construction process has to be analyzed in order to avoid problems during the construction phase. This includes the reachability of building elements and the positioning of construction robots. However, computing the reachability of target poses using inverse kinematics (IK) is time-consuming and does not allow for an immediate feedback. For this reason, a Random Forest classifier is trained to replace the calculation of IK. This decreases the computation time by a factor of more than 10^4 and allows for immediate feedback on reachability. It achieves a high accuracy of around 97%, which is shown for two exemplary robots. Due to the reduced computational effort, the classifier is used to optimize the robot base positions. The number of different base positions is minimized such that each target pose is reachable from at least one position. The optimization is successfully applied to a realistic pavilion with an accuracy of more than 99.99%. This contribution shows the immense advantage of using the proposed method to speed up planning of construction processes and enable co-design to improve and guarantee the feasibility of the construction process for a specific building design.

Keywords: Energy and Environment-Aware Automation, Mechatronics Systems, Building Automation
Abstract: Thermochemical energy storage systems are considered a loss free, environmentally friendly solution for storage of renewable energy. This work focuses on a specific application based on CaO/Ca(OH)2 for heat storage. During the endothermic dehydration of Ca(OH)2, heat is stored in the chemical bonds. By bringing CaO and H2O together in the exothermic hydration process, the stored heat is recovered. In this work, we set up a mathematical model to describe the dehydration process. The model covers the reaction driven by heaters in the reactor and the mass flow of gaseous water along a pressure gradient to a heat exchanger, where it condenses. Working towards model based control of the reactor system, the model equations are derived using enthalpy and mass balances as well as reaction kinetics. This yields dynamic equations for the reactor temperature, the water mass in the condenser and the conversion. Three unknown parameters related to heat transfer are identified via an optimization problem. For three data sets we apply a cross validation scheme, where two data sets are used for identification and the third one for model validation. All combinations of data sets in the cross validation were investigated and evaluated with the root mean square error and the qualitative model behavior. Although the identification of a distinct, globally optimal parameter set is challenging, the best parameter set is identified and can be used for future investigations and controller design.

Keywords: Mechatronics Systems, Automation Systems, Mechanism Design
Abstract: Robot technology has been integrated recently to streamline and alleviate the labor-intensive operations in convenience stores. We developed a specialized display shelf for automated product stocking by robots in convenience stores. The shelf includes a shelf–drawer mechanism and a product face-up mechanism. A transparent sheet (oriented polypropylene sheet) is installed around the shelf board, which rotates like a conveyor belt to move the products on the shelf board toward the front rail. This design enables the robot hand to operate without colliding with display fixtures. Experiments with rice balls, sandwiches, stick salads, and coleslaw confirmed that this method suits surface or line-contact products. However, the mechanism is less effective for curved packages or point-contact products because they might move out of the display lane or tip over. The developed display shelf was shown in the World Robot Summit Future Convenience Store Challenge @ IFAC 2023, and the team won the 1st prize.

Keywords: Logistics Systems, Formal Methods in System Integration, Software, Middleware and Programming Environments
Abstract: In this paper, we propose a traffic rule that can both improve the work efficiency of a logistics support robot performing picking work in a logistics center and avoid collisions at intersections, and describe the formal verification of a multiple AMR (Autonomous Mobile Robot) control system applying the proposed traffic rule using CSP (Communicating Sequential Processes) and time-expanded Timed-CSP. The proposed traffic rules include reducing interference between AMRs by considering their picking work positions on one-way aisle and prioritizing AMRs with less time remaining until the target work completion time, which can improve the overall work efficiency of multiple AMRs. The proposed system applying the traffic rules divides the intersection into three areas (intersection outer area, inner intersection area , and intersection center area) and sends status to other AMRs when entering or exiting the intersection, moving between areas, or stopping and resuming movement. Synchronous communication ports with the same behavior as CSP channels are used to send and receive data between AMRs. The results of the verification of the proposed traffic rules show that the Timed-CSP model of the proposed system does not cause deadlock or livelock, but the reachability validation of the CSP model without considering time shows that there is a possibility of collision at the intersection. When the CSP model is exhaustively verified, the event of movement may be executed before the event of sending and receiving statuses because time is not taken into account. Therefore, the verification results confirm that time-aware design is necessary for real-world systems. Formal verification results show that up to three statuses are sent when the inner intersection area moves, so the design must take into account the distance of the inner intersection area, the AMR's movement speed, and the communication delay time.

Keywords: Vision Systems, Human-Robot/System Interaction, Machine Learning
Abstract: Service robots operating in unfamiliar environments require capabilities for autonomous object recognition and learning from user interactions. However, present semantic segmentation methods, crucial for such tasks, often demand large datasets and costly annotations to achieve accurate inference. In addition, they cannot handle all possible objects or environmental variations without a large additional number of images and annotations. Therefore, this study introduces a learning system for semantic segmentation that combines 3D semantic mapping with interactions between an autonomous mobile robot and a user. We show that the proposed system can: 1) autonomously construct 3D semantic maps using an autonomous mobile robot, 2) improve the prediction accuracy of models pre-trained by supervised and weakly supervised learning in new environments, even without interaction, and 3) more accurately predict new classes of objects with a small number of additional coarse annotations obtained through interaction. Results obtained from experiments conducted in a real-world setting using models pre-trained on the NYU, VOC, and COCO datasets demonstrated an improvement in semantic segmentation accuracy when using our proposed system.

Keywords: Systems for Service/Assistive Applications, Automation Systems, Energy and Environment-Aware Automation
Abstract: Robotic systems are being used to automate convenience stores and retail stores, including competitions such as the World Robot Summit. Many of the systems that have been proposed so far are environmentally dependent, such as putting markers on product packages or using motorized shelves modified from existing shelves and are not easy to implement in stores. Therefore, we proposed a system to dispose of products with as few changes to the environment as possible. To realize the system, we proposed three approaches: markerless product disposal, autonomous movement with navigation, and customer detection with sensors that can be placed at any position. We conducted verification based on real convenience store scenarios and discussed the insights and challenges obtained from it.

Keywords: Decision Making Systems, Vision Systems, Machine Learning
Abstract: This study aims to build a system that bridges the gap between robotics and environmental understanding by integrating various foundation models. While current visual-language models (VLMs) and large language models (LLMs) have demonstrated robust capabilities in image recognition and language comprehension, challenges remain in integrating them into practical robotic applications. Therefore, we propose a visual-language decision (VLD) system that allows a robot to autonomously analyze its surroundings using three VLMs (CLIP, OFA, and PaddleOCR) to generate semantic information. This information is further processed using the GPT-3 LLM, which allows the robot to make judgments during autonomous navigation. The contribution is twofold: 1) We show that integrating CLIP, OFA, and PaddleOCR into a robotic system can generate task-critical information in unexplored environments; 2) We explore how to effectively use GPT-3 to match the results generated by specific VLMs and make navigation decisions based on environmental information. We also implement a photorealistic training environment using Isaac Sim to test and validate the proposed VLD system in simulation. Finally, we demonstrate VLD-based real-world navigation in an unexplored environment using a TurtleBot3 robot equipped with a lidar and an RGB camera.

Keywords: Vision Systems, Machine Learning
Abstract: In convenience store operations, there is a growing need for labor savings and automation. In this study, we focus on one of the most time-consuming tasks, product display, and develop a system to automate this task. In order for a robot to grasp a product, it is necessary to identify the product ID and estimate its position and posture. However, existing methods have difficulty identifying products in the same class, and no system with excellent occlusion avoidance and robustness has been established. In this study, we developed a product package with unique patterns and markers embedded in the package as a design. This package can be used for product recognition by deep learning. By learning the patterns and markers before the package is created, the learning can be performed efficiently. This paper describes the developed package and its recognition accuracy.

Keywords: Human-Robot/System Interaction, Human-Robot Cooperation/Collaboration, Human Interface
Abstract: People with visual impairments often rely on known reference points, navigational landmarks that they have identified, to move around indoors. However, it is difficult to identify landmarks using a white cane due to its short range and it not being able to provide information about visual characteristics of objects. We propose that a mobile robot equipped with a vision sensor can help people with visual impairments explore unfamiliar environments, identify navigational landmarks, and help guide them towards these landmarks, enabling them to better navigate independently in unfamiliar environments. The robot will operate in two modes, a wayfinding mode for identifying target objects, and a navigating mode to guide the user towards the identified target object. A prototype was developed and we performed two experiments to verify its operation. The prototype can easily switch between wayfinding and navigating modes with user intervention and can be used to guide a user to target objects around a room.

Keywords: Human-Robot/System Interaction, Motion and Path Planning, Hardware Platform
Abstract: This paper presents a solution for implementing a step-by-step teleoperation scheme on a humanoid robot NAO. The teleoperation scheme enables the user to manipulate the robot’s foot movements in a step-by-step manner using a pair of hand-held point devices with 3 degrees of freedom (DOFs) each. The controller for real robots suffers from errors resulting from factors such as material elasticity, motor performance, and sensor functionality. To solve the technical difficulties associated with implementing the scheme, we propose a solution that incorporates a virtual model and utilizes the divergent component of motion (DCM) feedback control. The DCM feedback control is realized by the center of mass (COM) admittance control and is enhanced by the angular momentum control. The teleoperated bipedal walking and fast swing of one leg experiments were conducted to validate the proposed solution.

Keywords: Human-Robot/System Interaction, Human-Robot Cooperation/Collaboration
Abstract: As the decisions we make become increasingly influenced by suggestions from Artificial Intelligence (AI), it is important to understand how the form and the accuracy of their suggestions affect our behavior. In this study, we investigated whether our decisions were affected by the accuracy of the AI used and the form of feedback. Ten participants evaluated whether facial images were real or fake under four conditions, namely, with suggestions given by a high or low accuracy AI and with or without the AI confidence level provided. The results suggest that, in the case of a high accuracy AI, the concordance rate between the decisions the participants made and the suggestions from AI were higher when confidence levels were provided compared to when the confidence levels were not show. However, in the case of a low accuracy AI, the concordance rate was lower when confidence levels were shown. Additionally, the accuracy of participant's answers increased when confidence level was provided, for both the high accuracy and low accuracy AI. In conclusion, providing confidence level information alongside AI suggestions may change the decisions made by human and the effects may be different depending on the confidence level of the AI.

Keywords: Human-Robot/System Interaction, Decision Making Systems, Machine Learning
Abstract: Service robots are designed to perform useful tasks for humans, which involve managing and combining a variety of skills in the form of global and local plans to solve any given task. In this work, we propose a framework to process natural language commands for general-purpose service robots where the robot should perform an arbitrary spoken command requested by a non-expert operator. Our system uses a Natural Language Processing (NLP) parser and a Conceptual Dependency (CD) builder to create CD structures that an expert system can employ to generate a global plan that the robot will execute. An extra simplification step using Large Language Models (LLM) was tested and evaluated in order to improve the accuracy of the system. Finally, our system has been tested in challenging environments in robot competitions and we have achieved promising results.

Keywords: Human-Robot/System Interaction, Human-Robot Cooperation/Collaboration
Abstract: The viewpoint position of motion parallax displays should ideally be aligned to the center of projection of the eye. However, most conventional motion-parallax displays have used image-based face tracking with a large delay instead of direct eye tracking, leading to the distortion of the presented image. This paper presents an eye-centered low-latency motion parallax display using fast and accurate viewpoint tracking. By incorporating a novel calibration method and the fastest monitor among the currently available ones, a motion-parallax display that operates at 360 Hz with a minimum delay of 3.78 ms from eye measurement to presentation is realized.

Keywords: Human-Robot/System Interaction, Micro/Nano Systems, Medical Systems
Abstract: In recent years, the demand for microinjections has increased manifold. The delivery of microinjections involves delicate manipulation based only on visual information, which requires a high amount of operational skill. To facilitate this process, we propose a 3D image-presentation micro-manipulation system with force feedback. By using a haptic device for manipulation, the proposed system allows the injection to be operated while receiving force feedback. We conducted manipulation experiments using porcine embryos to evaluate the effectiveness of the system.

Keywords: Integration Platform, Motion and Path Planning, Software Platform
Abstract: This paper presents a software system designed for indoor quadrupedal locomotion using vision-based techniques. It incorporates established technologies such as Simultaneous Localization and Mapping (SLAM), global planning, and composite Center of Gravity (COG) Trajectory planning methods to achieve both autonomous and semi-autonomous quadrupedal movement. The software solution offers a comprehensive approach to locomotion challenges, encompassing vision-based foothold planning and obstacle avoidance. With a modular architecture that supports diverse controller implementations, the software demonstrates its versatility. Overall, this research contributes a comprehensive software framework for indoor quadrupedal locomotion, exhibiting promising results in simulation and offering potential for real-world deployment.

Keywords: Integration Platform, Software, Middleware and Programming Environments, Network Systems
Abstract: Machine learning is one of the most promosing technologies to improve predictive maintenance especially in industrialized settings. However, oftentimes, companies are lim- ited in their ability to generate and to share machine data, which this in turn limits the amount of such. Therefore, contemporary systems’ prediction of failures and necessary maintenance all too often rely on manually set thresholds and hand-designed heuristics. On this account, this paper proposes an end-to-end federated-learning-based system to address these challenges in a real-world industrial setting by sharing data and training progresses across multiple entities that contribute to a global prediction model. Thereby, our approach combines federated learning with edge computing to leverages the advantages of sharing computing resources and utilizes parameters of locally trained models (rather than actual machine data). We integrated the approach into an industrial system for predictive maintenance of crucial parameters of industrial robotic grippers and compared it against reference models conventionally trained on one source of data. Our findings indicate that limited datasets and models created at individual companies yield — once aggregated into a global ML model results superior to heuristic approaches and better results than individual companies’ mod- els. This work further shows that edge computing provides a suitable infrastructure for federated learning systems.

Keywords: Integration Platform, Soft Robotics, Network Systems
Abstract: Web of Tactile Things (WoTT) aims toward a standardized platform designed for tactile devices, facilitating the exchange and updating of data across diverse domain web services. However, the requirements for specified devices pose a challenge when deploying on low-profile devices. Furthermore, the concentration of processing the sensing and server operations on a single computer creates a bottleneck that restricts the system's scalability. To enhance the potential applications, this study introduces an extended platform based on WoTT that decentralizes processing. This separation involves distinct processing for the tactile sensor and the WoTT server becomes independent elements. The tactile sensor is integrated with an onboard computer for sensing tasks, enabling wireless transmission of tactile data packets. In this proposed system, each external device can establish wireless connections with the server via the MQTT (Message Queueing Telemetry Transport) message protocol. In order to validate this approach, we construct a comprehensive system incorporating tactile sensors, AR devices, and 2D linear stages. Leveraging a deep neural network (DNN), the tactile sensor's data drives the anticipation of soft skin deformation by capturing visual cues through marker displacement. Tactile data is then encoded and transmitted to end devices using the MQTT protocol, directing the 2D linear stage's interaction with a phantom arm, while AR devices visualize the tactile sensor's soft skin deformation.

Keywords: Integration Platform, Hardware Platform, Automation Systems
Abstract: We constructed a camera relay system using a motor equipped with a mechanical brake to reduce the constant micromovement caused by the noise control system of a mirror-driven active vision system that enables high-speed magnified tracking imaging and measures the micro-vibrations of a moving object. We verified that the proposed motor equipped with a mechanical brake achieved a noise reduction that was approximately five times higher than that achieved by a conventional motor without a mechanical brake. As a result, we measured 0.01~pixels, a 0.08~mm magnitude of micro-vibrations in a wide range of lateral direction with respect to 0.16 m x 0.09 m of the camera-field angle.

Keywords: Integration Platform, Autonomous Vehicle Navigation, Software Platform
Abstract: This video demonstration paper presents an integrated adaptive control system for obstacle detection and online speed adaptation of autonomous drones. It addresses the challenges of controlling drones for obstacle avoidance. The results demonstrate that the control system allows our drone to detect obstacles using LIDAR sensors. It can learn autonomously in real-time, adjusting its flying speed as it approaches obstacles to ensure a safe flight. Furthermore, the drone can deal with unexpected static and moving obstacles in different outdoor environments, both during the day and at night.

Keywords: Integration Platform, Mechanism Design, Automation Systems
Abstract: In the world of table tennis, practices often involve the use of numerous balls, making post-practice collection a tedious and time-consuming task. To address this challenge, we presented an autonomous robot tailored for table tennis ball collection. This robot features a differential-drive wheeled platform, a unique ball collection mechanism using rotating elastic blades, an on-board RGB camera, and a single board computer for data processing. Balls are detected using a deep learning algorithm, allowing the robot to autonomously navigate the nearest ball and collect it. Under a consistently controlled environment include a limited field size and a set number of balls, experiments were conducted to clearly evaluate the robot's core functions and performance. Experimental results revealed that despite employing a straightforward operational algorithm, the robot demonstrated performance comparable to human-operated remote collection in terms of speed and accuracy. Importantly, this study uniquely focuses not on maximizing efficiency but rather on establishing a baseline for future research in this area, setting a benchmark for future studies. Furthermore, our findings provide the foundation for enhancing ball-collecting robots in table tennis settings.

Keywords: Biologically-Inspired Robotic Systems
Abstract: Recently, robots that operate in unknown environments, such as rescue robot, agriculture robot, construction robot, etc., have been attracting attention. To adapt to unknown environments, multi-legged robots are advantageous because of their high mobility and stability. However, they have the problem of slow walking speed. In this study, we have proposed a multi-legged robot that can move in two ways: centipede-like walking and cylindrical rolling. On an uneven terrain, such as rubble, narrow areas, and steps, the robot moves by performing centipede-like walking, whereas on flat terrains, such as roads and pass-ways, it moves quickly, by rotating its cylindrical body. In this article, we describe the developed prototype and the experiments conducted on it to evaluate its performance.

Keywords: Biomimetics, Mechatronics Systems, Mechanism Design
Abstract: Movements of knee joints are relative motions between tibia and femur, which include rolling and sliding. Conventional wearable knee assistive device utilized a hinge joint and generates nonnegligible mismatched motions under deep flexions. To assist the knee motion including the deep flexion, we have developed several types of the biomimetic knee joint (BKJ) mechanism that mimics the rotational centers of biological knees. A polycentric BKJ which is a polycentric joint structure that mimics the rollback motions of knee joints by combining two gears with different radii is used in this study. From a viewpoint of the complexity of the device, the passive devices have a significant advantage. However, a lack of the automatic adjustment function is a big disadvantage. To enhance the usability and applicational field, we propose the Semi-active Knee Joint (SKJ) in this study. The SKJ has a polycentric gear mechanism and linear motion springs. In addition, the spring constant is adjustable from a weak spring to a strong spring. The basic mechanism, design method, and design processes are reported in this paper. As a result of an optimal design of the gear system based on the measurement result of a knee joint motion for a subject, we adequately designed a pair of gears for a tailer made SKJ with small misalignment of less than 5 mm even in the deep flexion over 90 degrees. In addition, non-linear function of the assistive torque was analyzed for a mechanical model. As a result, spring constant, k is decided to be 45 N/mm for the upper knee and 20 N/mm for the lower knee.

Keywords: Biomimetics, Biologically-Inspired Robotic Systems, Soft Robotics
Abstract: Snake-like robots have attracted attention as robots that can travel over rough terrain, such as disaster sites and planetary exploration applications, whereas wheeled mobility mechanisms cannot. However, in the previous studies on snake-like robots, the issue of power consumption due to driving a large number of actuators remained. In this study, we propose a mechanism that transforms the backward link of a snake-like robot into a wheel-like shape to realize a three-wheeled vehicle mode on flat ground that provides the same reliability, traveling speed, and efficiency of traveling as a wheeled mobile robot. However, to improve travel efficiency and running stability, the motion in the undulation and wheel modes must be properly designed. Therefore, we designed motions in these modes and constructed a method for switching between them. Experiments were conducted to verify the travel efficiency of the proposed wheeled mode, and it was shown that, it can achieve greater travel efficiency than undulation.

Keywords: Biomimetics, Biologically-Inspired Robotic Systems, Soft Robotics
Abstract: In this study, we attempt to promote fermentation using a device that simulates intestinal peristalsis. Fermentation by beneficial bacteria maintains the intestinal environment. Moreover, there is a possibility that the acceleration of fermentation is related to intestinal movements such as peristalsis. Therefore, we aim to elucidate the relationship between fermentation and peristalsis, and further promote fermentation using a device that simulates peristalsis. In this paper, as an initial study, we used the device to confirm the progress of fermentation using a substance with a large viscosity. The results revealed that, the higher the viscosity, the more stable the degree of mixing and the greater the progression of fermentation.

Keywords: Biomimetics, Biologically-Inspired Robotic Systems, Automation Systems
Abstract: In recent years, robots that move on tubular structures have attracted much attention. These robots are used in a variety of applications and provide very significant benefits such as increased work efficiency, improved safety, and reduced work costs. Robots that move on tubular structures include robots that move while grasping objects, robots that move while attached by suction cups or adhesive materials, and robots that move by wheels. However, conventional robots have large, rigid parts and cannot move in confined environments because they need sufficient mechanical strength to climb against gravity. This is where our research group’s Funabot- Finger Cot comes in. This device is a small device with pneumatically expandable artificial muscles fixed to a cloth, and its movement can be controlled by controlling the pattern of air pressure application. The repeated contraction and relaxation characteristics of the device suggest that it could be used as a robot that moves on a tubular structure that mimics the peristaltic movement of living organisms. Peristaltic motion is the coordinated contraction and relaxation of muscles, which allows the robot to move over the ground, gaps, and objects. The Funabot-Finger Cot makes several contributions. First, because it is made of flexible fabric, it can be wrapped around structures of various shapes. Its thinness also allows it to move efficiently in constrained environments that are difficult for conventional robots. Secondly Cost-effective and cost-saving. Because this is small robot and made of common fabric and a small amount of artificial muscle, they can be made inexpensively and easily. The most notable feature is its versatility. Finally, the fact that the actuator supports both circular longitudinal contraction allows for more complex movements. This study will focus on the third advantage in particular, evaluating movement over tubular structures in circular contraction and in longitudinal contraction.

Keywords: Software Platform, Hardware Platform, Software, Middleware and Programming Environments
Abstract: In this paper, a parallel computing solution has been proposed to improve the timing of channel estimation (CE) in 5G applications. By utilizing the parallelism scheme of the digital signal processor with the Single Instruction Multiple Data (SIMD) architecture, the total computation time is drastically reduced to 0.3ms instead of over 8.5ms with a scalar non-SIMD architecture based on the same hardware platform with the proposed SIMD vDSP, and both are at the same 500MHz operating frequency. Performance comparisons have been conducted to highlight the efficiency of the parallelism scheme with SIMD. Channel estimation schemes including least squared (LS), interpolation, and noise calculation have been implemented with the proposed parallel computing architecture for demonstration.

Keywords: Software Platform, Sensor Fusion, System Simulation
Abstract: This paper reports on fault detection and fault-tolerant control for the safe operation of water hydraulic robots equipped with Air-Hydraulic Servo Boosters (AHSB). By maximally utilizing the redundant sensors embedded in AHSB, model-based algorithms of failure detection, isolation and identification are implemented using Kalman filter and disturbance observer. Simulation results show that a single-joint robot arm can continue position-tracking tasks against sudden disconnection of sensors, internal friction fluctuation, and external torque application, hence demonstrating the effectiveness of the proposed method.

Keywords: Software, Middleware and Programming Environments, Machine Learning, System Simulation
Abstract: Maintaining stability in bipedal walking remains a significant challenge in humanoid robotics, largely due to the numerous involved hyperparameters. Traditional methods for determining these hyperparameters, such as heuristic approaches, can be both time-consuming and potentially suboptimal. In this paper, we present an approach aimed at enhancing the stability of bipedal gait, particularly when faced with floor perturbations and speed variations. Our main contribution is the integration of intrinsically stable model predictive control (IS-MPC) and whole-body admittance control within a closed-loop reinforcement learning system. We devised a reinforcement learning plugin, implemented in the {tt mc_rtc} framework, that allows the control system to continuously monitor the robot's current states, maintain recursive feasibility, and optimize parameters in real-time. Furthermore, we propose a reward function derived from a combination of changes in single and double support time, postural recovery, divergent control of motion, and action generation grounded in training optimization. In the course of this research, we conducted experiments on a real humanoid robot to validate initial aspects of our work. The integrated module's effectiveness was further assessed through comprehensive simulations.

Keywords: System Simulation, Systems for Service/Assistive Applications
Abstract: Fire preparedness is an essential issue in protecting both human life and property. In addition to the fact that, especially in urban areas, buildings tend to be more densely built, the predominant building structure in many countries is still wood, which is vulnerable to fire. As a result, large-scale building fires, known as ”urban fires”, have occurred many times. Because fire is a very complex and uncertain event, a variety of study approaches have been taken. In this paper, we focus on the impact of wind speed and direction on urban fires and develop a fire model that can take advantage of realtime weather information using a graph network. The impact of the accuracy of wind speed and direction information on the accuracy of the fire model will also be verified through the simulation based on an actual fire using the developed fire model. The fire model constructed in this paper assumes urban fires that cause large-scale damage among building fires, and the target of prediction is the damage to buildings. A new contribution of this paper is the identification of the impact of wind on fire model accuracy and the proposal for highly accurate fire damage prediction with real-time weather information.

Keywords: Systems for Service/Assistive Applications, Human-Robot/System Interaction, Path Planning for Multiple Mobile Robots or Agents
Abstract: The Future Convenience Store Challenge (FCSC) at the World Robot Summit (WRS) includes a Customer Interaction Task that proposes new customer service services using robot technology in a convenience store in the near future. Conventional proposals have predominantly focused on providing services in specific situations or utilizing request-based approaches. However, the conditions within convenience stores vary throughout the day, with changes in customer congestion and product stock levels occurring due to different time periods. Therefore, in this study, we have proposed several new services that align with the flow of a convenience store's day, aiming to alleviate the burden on store clerks and save labor through situational collaboration between humans and robotic systems.

Keywords: Virtual Reality and Interfaces, Vision Systems, Human-Robot Cooperation/Collaboration
Abstract: Robot manipulation in retail environments is a challenging task due to the need for large amounts of annotated data for accurate 6D-pose estimation of items. Onsite data collection, additional manual annotation, and model fine-tuning are often required when deploying robots in new environments, as varying lighting conditions, clutter, and occlusions can significantly diminish performance. Therefore, we propose a system to annotate the 6D pose of items using mixed reality (MR) to enhance the robustness of robot manipulation in retail environments. Our main contribution is a system that can display 6D-pose estimation results of a trained model from multiple perspectives in MR, and enable onsite (re-)annotation of incorrectly inferred item poses using hand gestures. The proposed system is compared to a PC-based annotation system using a mouse and the robot camera's point cloud in an extensive quantitative experiment. Our experimental results indicate that MR can increase the accuracy of pose annotation, especially by reducing position errors.

Keywords: Software, Middleware and Programming Environments, Integration Platform, Software Platform
Abstract: Various robot middleware platforms are used to build robot systems. Many modules developed by a middleware platform can only be used on the middleware, and it is difficult to interoperate between different middleware. Therefore, it is necessary to develop a new module according to the middleware to be used, which leads to an increase in development time and, as a result, an increase in product cost. By creating a framework that enables the interoperability of software modules between different middleware platforms, we aim to improve the reusability of software modules and operate without awareness of the existence of middleware platforms even when users operate multiple middleware platforms on the same system. We focused on wasanbon, a framework for the development of OpenRTM-aist. We incorporated into wasanbon a module that converts data between OpenRTM-aist and ROS and a system management and operation method based on abstracted commands, which made it possible to operate a system that can operate OpenRTM-aist and ROS. We created an in-hospital guidance system using this framework and confirmed that the system can be operated by users without related knowledge, even when the system contains several software modules with two types of middleware platforms.

Keywords: Control Theory and Technology, System Simulation, Decision Making Systems
Abstract: This paper presents a control algorithm based on the Evolving Type 2 Quantum Fuzzy Neural Network (ET2QFNN) for characterizing and controlling the joint-space motion of a manipulator. The approach integrates ET2QFNN, a hybrid neural network model that combines principles from quantum fuzzy sets and neural networks, to address the inherent complexity and uncertainty associated with manipulator systems. Unlike traditional control methods, the ET2QFNNbased approach learns and adapts through simulations of closed-loop torque control, incorporating joint position commands and corresponding motion data. By utilizing data from previous cycles, the ET2QFNN enhances the manipulator’s performance and stability, particularly in repetitive tasks. To assess its effectiveness, a comparative analysis is conducted, contrasting the approach with finely tuned proportional-derivative (PD) control and integral sliding mode control (iSMC) in scenarios with and without non-parametric uncertainties. The results demonstrate the unique advantages and superior performance of the ET2QFNN-based control algorithm in addressing these challenges

Keywords: Control Theory and Technology, System Simulation, Mechatronics Systems
Abstract: A compact propulsion system that can operate in a low-power and low-temperature environment is necessary to expand the field of activity of CubeSats from Earth orbit to deep space. Tohoku University, Hokkaido University, and Muroran Institute of Technology are developing the 3U CubeSat ”HOKUSHIN-1” for future lunar exploration by CubeSats. For orbit maintenance and 3-axis attitude control, this satellite is equipped with a new 1U size cold gas propulsion system developed by Patchedconics, LLC, and is planned for technical demonstration on the ISS orbit. This paper proposes a system design and operation method to prevent hazards caused by malfunction of the propulsion system and to ensure safety during orbit operations. Furthermore, the attitude and orbit control capability of the new propulsion system is evaluated by numerical simulation.

Keywords: Medical Systems, Formal Methods in System Integration, Human-Robot/System Interaction
Abstract: In response to the growing demand for precise and affordable solutions for Image-Guided Spine Surgery (IGSS), this paper presents a comprehensive development of a Robot-Assisted and Navigation-Guided IGSS System. The endeavor involves integrating cutting-edge technologies to attain the required surgical precision and limit user radiation exposure, thereby addressing the limitations of manual surgical methods. We propose an IGSS workflow and system architecture employing a hybrid-layered approach, combining modular and integrated system architectures in distinctive layers to develop an affordable system for seamless integration, scalability, and reconfigurability. We developed and integrated the system and extensively tested it on phantoms and cadavers. The proposed system's accuracy using navigation guidance is 1.02±0.34 mm, and robot assistance is 1.11±0.49 mm on phantoms. Observing a similar performance in cadaveric validation where 84% of screw placements were grade A, 10% were grade B using navigation guidance, 90% were grade A, and 10% were grade B using robot assistance as per the Gertzbein-Robbins scale, proving its efficacy for an IGSS. The evaluated performance is adequate for an IGSS and at par with the existing systems in literature and those commercially available. The user radiation is lower than in the literature, given that the system requires only an average of 3 C-Arm images per pedicle screw placement and verification.

Keywords: Human-Robot/System Interaction
Abstract: This paper proposes hip flexion/extension assistance measures to improve dynamic gait stability. We conducted an experiment with the elderly to show that hip flexion/extension assistance at the moment of initial foot contact improves dynamic gait stability. The proposed measures were verified by developing a new system with the assistive powered wear HIMICO. The stability of the target was improved, and the gait characteristics were analyzed to propose walking-assist measures.

Keywords: Human-Robot/System Interaction, Modeling and Simulating Humans, Control Theory and Technology
Abstract: To alleviate the lack of workforce endemic in contemporary societies, inventory management can be delegated to robots. However, current robotic solutions cannot operate effectively in narrow spaces that have limited unoccupied pathways, such as storage rooms. Humanoid robots can work in facilities designed for humans, including narrow spaces with partially obstructed pathways, as they can overcome small obstacles on the floor by stepping over them. However, their limited torque makes it difficult for the robots to carry many of the objects typically found in storage rooms such as boxes. On the other hand, when holding an object, using multiple body parts can help to distribute the object's load. In this research, we propose a holding strategy for humanoid robots using both arms in conjunction with the torso to prevent exceeding the joint torque limit. We formulate the hands' force and the coordinates of the robot's center of gravity when holding an object with and without using the torso. Using the humanoid robot HRP-4, we compare the torque on the robot's joints when holding an object with two and three contact points. The results show that the three-point holding reduces the torque load on the shoulder and the chest joints. Moreover, we measure the torque of all joints in various torso-bending/leaning and arm-up/down postures while holding a 2-kg object using both arms and the torso. The results show that the leaning torso posture using the chest joint reduces the torque load on the shoulder and chest joints.

Keywords: Human-Robot/System Interaction, Human-Robot Cooperation/Collaboration, Machine Learning
Abstract: This research investigates the intricacies of women's perceptions and experiences when interacting with a robot trained on imbalanced user motion data, harnessed from wearable Inertial Measurement Unit (IMU) sensors. Utilizing motion data as the primary communication conduit between user and robot, the study provides a multifaceted exploration into human-robot collaboration dynamics. A cross-gender comparison reveals disparities in performance outcomes, highlighting a discernible difference between male and female users. Furthermore, our study delves deep into the stages of trust building, its subsequent violation, and the repair process, offering insights into how these phases distinctly influence women's experiences. A pivotal aspect of our research culminates in unveiling the impacts of awareness. When informed of the robot's biased nature, many women participants exhibited unexpected responses in attributing the subpar performance. These findings illuminate the subtle, yet profound, implications of gender biases in robotic training data sets and underscore the imperative need for balanced and transparent Artificial Intelligence (AI) and robotic systems.

Keywords: Human-Robot/System Interaction, Soft Robotics, Intelligent and Flexible Manufacturing
Abstract: Labor shortages and low productivity demand more collaborative operation systems using industrial robots in the current global manufacturing industry. Injuries will possibly occur due to the closer interaction between the robot system and humans. In the field of skin contact safety, the techniques and methods of quantitative assessment are expected. A human dummy consisting of built-in sensors is a solution to obtain the internal physical field. In this study, A dummy finger equipped with a built-in film sensor for contact safety assessment is proposed and evaluated by experiment. The strain measurement model and strain energy sensing principle, which are the most general assessment indexes, are explained. Two kinds of contact conditions on the finger dummy, press and press with slip, are conducted for evaluation. Shear strain and vibration are both successfully measured and reflected in the signal. Measurement of the strain energy transformed from the strain is validated by comparing it with the work result obtained from the force sensor. This sensor system combined with the human body model is evaluated as potentially feasible for assessing physical contact safety.

Keywords: Human-Robot/System Interaction, Soft Robotics, Vision Systems
Abstract: We propose a human–machine system that provides force feedback coaching for forehand swings in table tennis. In the proposed system, a fabric actuator suit with artificial muscles provides motion sensation to the operator. Furthermore, real-time coaching is achieved by estimating the future arm angles using artificial intelligence. The proposed coaching system is tested experimentally.

Keywords: Human-Robot/System Interaction, Human Interface, Human Factors and Human-in-the-Loop
Abstract: Before the existence of computers, dialogue was the primary format of communication between humans. However, with the advancement of Large Language Models (LLMs), communication has shifted to using keyboards or displays. As LLM technology continues to advance, new methods of communication must be developed to accommodate it. In our research, we created two LLM agents and experimented with the ability of users to intervene in a conversation between the agents using an EEG device. The results demonstrated that communication with LLMs can be achieved without the need for keyboards or displays.

Keywords: Systems for Field Applications, Human-Robot/System Interaction, Human Interface
Abstract: Most existing retrofit remote-control systems for backhoes are large, hard-to-install, and expensive. Therefore, we propose an easy-to-install and inexpensive teleoperation system. The proposed system comprised remote-control and sensing systems. The remote-control system retrofits robot arm-based devices to ”Cat Command”, a compact embedded teleoperation system with a limited communication range, and controls these devices via a 5G commercial network to realize control from a remote office. Because this system does not require any additional modifications to the embedded control unit in the cockpit, the operator can continue working in the cockpit even if the backhoe is remotely controlled. The system enables the remote control of various devices from an extremely long distance by changing the joint parts between the robot arm and the embedded remote-control device. The sensing system estimates the posture and position of the backhoe by attaching original sensing devices to the backhoe. In addition, a 360◦ camera was installed in the cockpit to transmit work images from the construction site to a remote office in real time. The sensing device was smaller and lighter than conventional devices. We confirmed that the proposed system can be used to operate a construction site backhoe from a remote office, and that the system can be used to excavate soil using an actual backhoe.

Keywords: Systems for Field Applications, Environment / Ecological Systems, Mechanism Design
Abstract: Marine microplastics originate from plastic products and are crushed as they drift through the ocean, posing a serious threat to marine ecosystems. Nonetheless, collecting these scattered and small microplastics by hand from washedup beaches is challenging. Consequently, we are working on designing a cleaning robot to automatically gather marine microplastics on beaches. This paper outlines the primary mechanism of the cleaning robot and explores the necessary functions for effective microplastic collection. Two kinds of experiments exploring the behavior of sand in stationary and dynamically states, ”Pseudo-Angle of Repose Measurement Experiment” and ”Claw Excavation Experiment”, are carried out using an experimental model of a cleaning robot.

Keywords: Systems for Field Applications, Mechanism Design, Automation Systems
Abstract: The decline in agricultural productivity due to climate change, diseases, and pests has raised concerns for farmers globally. To address these challenges, innovative approaches integrating mobile robotics into agriculture have emerged to assist human labor. This study presents a novel solution utilizing a Delta robot with anisotropically flexible structure for nondestructive data collection from plants. Three data-gathering methods are introduced utilizing the platform. Experimental results demonstrate the advantages of the proposed approach in capturing intricate plant details and inspecting concealed regions, particularly beneficial for detecting pathogens and pests. This proposed solution has the potential to offer greater detail than aerial vehicles, while also introducing a new way of monitoring plant health under foliage. This research contributes to enhancing plant health monitoring in agriculture through agile robotic systems.

Keywords: Systems for Field Applications, Integration Platform, Vision Systems
Abstract: This paper proposes a control method for autonomous forklifts to unload pallets on inclined surfaces, enabling the fork to be withdrawn without dragging the pallets. The proposed method applies the Iterative Closest Point (ICP) algorithm to point clouds measured from the upper region of the pallet and thereby tracks the relative position and attitude angle difference between the pallet and the fork during the unloading operation in real-time. According to the tracking result, the fork is aligned parallel to the target surface. After the fork is aligned, it is possible to complete the unloading process by withdrawing the fork along the tilt, preventing any dragging of the pallet. The effectiveness of the proposed method is verified through dynamic simulations and experiments using a real forklift that replicate unloading operations onto the inclined bed of a truck.

Keywords: Systems for Field Applications, Autonomous Vehicle Navigation, Automation Systems
Abstract: The introduction of unmanned underwater vehicles (UUVs) for the inspection of underwater structures is progressing rapidly. However, their control is difficult because they are subjected to several disturbances and uncertainties. Among the UUVs, the tethered types are significantly more difficult to control because tether tension also acts on the vehicle. Hence, this study aims to clarify the factors lurking in the tether that make control difficult and contribute to achieving excellent control performance of tethered underwater vehicles. We attached two positively buoyant tethers of different stiffness and material to an actual underwater vehicle platform. We then conducte an experiment on a breakwater to examine if different tether specifications make a difference in control behaviour. Our experimental results show that one tether has little impact on control, while the other tether reduces control accuracy. Consequently, we derive key factors that can have a significant impact on the control applications, as well as ocean currents and thruster specifications.

Keywords: Systems for Field Applications, Integration Platform, Automation Systems
Abstract: Performance evaluation of rotary blades on cultivators is mainly conducted by performing tests on actual equipment. However, issues, such as the high cost of prototype blades and the difficulty in precise measurement, exist. Therefore, attempts are being made to use the individual element method for the evaluation of tilling operations. However, the determination of its parameters requires comparison with experimental results based on physical phenomena of soil in cultivation operations. This study aims to construct a scaled model experimental method in which torque and soil dispersal conditions can be sensed. Specifically, a 1/4-scale model experiment apparatus was developed to measure torque and three-component force, which are loads during cultivation, as well as soil dispersal, which is an index for evaluating cultivation performance.

Keywords: Mechanism Design
Abstract: 3D printers capable of using in various fields, and study on toughness and molding accuracy of 3D printed parts are being conducted. However, there has been almost no study on changes in size, mass, and toughness due to hygroscopicity. In this study, changes in size and mass of 3D printed parts that absorbed moisture have been measured in long term. Based on these results, analyzed the tendency of size and mass changes and shown the tendency of size shrinkage to change depending on the amount of moisture in the 3D printed parts. It has also been confirmed the toughness changes due to moisture absorption by three-point bending test. Characteristics of size and mass changes due to infill structure differences has been confirmed. Based on the results of these experiments, the dependence of changes in size, mass, and toughness due to moisture absorption on the amount of moisture in molding materials and the environment in which they are used are discussed.

Keywords: Mechanism Design, Mechatronics Systems, Hardware Platform
Abstract: In recent years, the demand for industrial robots has increased and is expected to continue to grow. In addition, efforts toward the United Nations’ Sustainable Development Goals and carbon neutrality are gaining momentum worldwide. The purpose of this study is to improve the energy efficiency of robots, and this paper investigates materials for lightweight reduction gears. To reduce robot weight, it is necessary to replace conventional metal materials with lightweight materials such as composite materials and resins. However, in general, weight reduction tends to reduce stiffness, resulting in a decrease in the speed and precision performance required for industrial robots. We have evaluated the applicability of replacing metal parts with plastic materials made using a fused deposition modeling 3D printer. In the previous paper, only metal parts and 3D printed resin parts were compared, but in this paper, machined CFRP and POM parts, and 3D printed resin parts with increased infill ratio are added to the material types and compared. The results show that there are differences in mass, no-load torque, stiffness, efficiency, etc., depending on the combination of materials, and that weight reduction can be achieved with minimal performance degradation by selecting a combination appropriate for the application.

Keywords: Mechanism Design, Hardware Platform, Biologically-Inspired Robotic Systems
Abstract: We have been continuously research and developing new robots for the purpose of rice paddy weed control. During the period of one month following rice planting, utilizing a robot that moves within the water paddy field to stir and cloud the water can lead to effective weed control by inhibiting weed photosynthesis. However, paddy fields contain suspended matter such as floating water plants, and there are obstacles like growing rice plants, making it impossible to use existing mobile robots. To address this issue, we are developing a new snake robot capable of swimming on the water surface to move. In this paper, we present the development of a snake robot and report field tests conducted in the real environments.

Keywords: Micro/Nano Systems, Systems for Field Applications, Control Theory and Technology
Abstract: In industrial fields such as manufacturing and inspection equipment for electronic devices, micro-vibrations need to be suppressed as they deteriorate the performance of equipment. We previously proposed an add-on type active vibration suppression system as a reliable and simple solution for vibrations that affect electron microscopes, which is an application example of a nanometer-order system. The system suppresses nanometer-order vibrations in the structure simulating an electron microscope by controlling the output of high-thrust piezoelectric actuators using acceleration sensors. However, if the output of the actuator exceeds a maximum limit due to large vibrations, the active vibration suppression system causes a high-frequency disturbance that deteriorates equipment performance. In this paper, we propose a multi-scale vibration control method which estimates the amplitude of the controller output in real time and adjusts the feedback gain on the basis of the estimated amplitude. The proposed method can suppress a wide range of vibrations from nanometer- to micrometer-order, thereby suppressing vibration and stabilizing the active vibration suppression system. The vibration suppression performance of the proposed system is evaluated through experiments using a structure simulating an electron microscope.

Keywords: Micro/Nano Systems, Automation at Micro-Nano Scales
Abstract: This paper reports on the operation and immunoassay performance of a disposable polystyrene microfluidic device manufactured by injection molding. The device is a centrifugal microfluidic device equipped with our originally developed Control of Liquid Operation by Centrifugal hydro-Kinetic(CLOCK). This allows complex chemical processes consisting of multi-step unit operations to be carried out simply by spinning at a constant speed. Polystyrene has excellent adsorption properties for antibodies and is commonly used as a base material for immunoassay methods and cell culture. Therefore, realizing a device made of polystyrene is expected to increase the amount of adsorbed antibodies and the sensitivity of the measurement system. The device can realize a liquid operation to perform sandwich ELISA and it has been confirmed that it can perform automatic ELISA with stable operation. In the evaluation of the assay system using mouse-derived IgG as antigen, a calibration curve was successfully measured, with a lower limit of detection (LOD) of 0.53 ng/ml when 1% BSA was used as solvent and 6.67 ng/ml when the supernatant obtained by centrifugation of sterile blood from sheep was used as solvent.

Keywords: Control Theory and Technology, Sensor Fusion, System Simulation
Abstract: Tohoku University and Chiba Institute of Technology have been working on the development of a Micro Space Telescope satellite for planetary science exploration. Long-term photometric observations are indispensable in the search for exoplanets with long orbital periods using the transit method. Improvement of the detection rate needs miniaturizing and improving satellite systems while conducting coordinated observations. Image blurring adversely affects photometric accuracy, thereby diminishing the likelihood of discovering exoplanets, and hence, accurate pointing control is imperative for telescopes. This study aims to achieve arcsecond-level pointing accuracy by using MEMS mirror tip-tilt control to rectify the optical path, well better than what can be achieved with traditionally attitude control on the satellite itself. This research will encompass both laboratory experiments employing the star simulator and simulation outcomes pertaining to MEMS mirror control and satellite attitude control.

Keywords: Autonomous Vehicle Navigation, Factory Automation, Sensor Fusion
Abstract: Environmental changes can severely disrupt the docking sections of transportation systems in factories. Such disruptions not only halt the production line but also necessitate human intervention for loading and unloading, posing safety risks. To address this, we introduce an autonomous docking system designed for resilience in semi-unstructured environments, especially when faced with varying light conditions and physical alterations. Our solution involves an advanced AruCo marker system. By integrating the yaw orientation of the marker with angular velocity data from an inertial measurement unit, we achieve a homogeneous matrix reconstruction. This enhanced marker data then guides a reference trajectory for docking, controlled by a PI Pure Pursuit mechanism. To further improve marker recognition, we have incorporated reflective and anti-reflective materials and modulated the marker's white margin. Our test results indicate a significant improvement in the detection range of a 21cm marker expanded from 60cm to 2m. Moreover, our system ensures a docking precision of ±31mm and ±1.82°.

Keywords: Intelligent Transportation Systems
Abstract: Mitigating the challenge of efficiently estimating daily dynamic travel demand is crucial for agent-based traffic simulations and on-demand transport services. However, existing methods relying on minimization of objective loss function are time-consuming, primarily due to the considerable volume of spatiotemporal observational data and the multitude of variables that necessitate optimization. In this study, we propose a hierarchical approach that employs a coarse-to-fine temporalresolution strategy to address this time-consuming issue of the estimation process. Our method leverages a gradient descent-based optimization technique that minimizes vehicle count errors between simulated and observed values at each temporal resolution layer with different granularity. We effectively reduced the time consumption at finer temporal resolutions by initially estimating travel demand using a coarse temporal resolution configuration and employing the corresponding results as an initial solution for finer temporal resolution layers. This architectural design, comprising multiple temporal resolution layers and the propagation of the estimated travel demand from top to bottom, considerably reduces overall estimation time. To evaluate our approach, we conducted experiments that utilize an actual large-scale road network and commercial GPS data collected from probe vehicles. Furthermore, we compare the time consumption of our proposed method, which employs multiple temporal resolutions, with a method that utilizes a single fine temporal resolution. The experimental results reveal that the proposed approach substantially reduces travel demand estimation time by 52%.

Keywords: Intelligent Transportation Systems, Autonomous Vehicle Navigation, Human-Robot/System Interaction
Abstract: This paper proposes an automatic berthing system that performs berthing according to the skill of the operator. To make a maneuvering plan according to the navigator’s skill, AIS data on the maneuvering method of berthing in actual seas was analyzed, and three maneuvering plans were derived. The proposed system uses a deep neural network as a controller to execute the ship berthing. Simulation experiments show the effectiveness of the proposed system.

Keywords: Intelligent Transportation Systems, Formal Methods in System Integration, Mechatronics Systems
Abstract: This paper explores the integration of model-based systems engineering (MBSE) and scenario-based engineering (SE) approaches in the development of advanced driver assistance systems (ADAS) and automated driving (AD). The increasing complexity of automotive systems necessitates efficient system design and validation processes. MBSE provides a framework for capturing system functionality and managing complexity, while SE focuses on testing and analyzing system features across various scenarios. By combining these approaches, system development efforts can be streamlined, leading to more efficient project management. This paper reviews existing literature on MBSE and SE in the context of ADAS development and highlights the potential benefits of their integration. An MBSE workflow is introduced where the key focus is on the behavioral aspects of the system. Additionally, a scenario generation pipeline which constitutes scenario databases is proposed to ensure comprehensive testing of system requirements. Results show that with automated scenario generation, the target vehicle cut-out scenario is successfully realized and simulated based on the formally described system behavior as part of the MBSE model.