Keywords: Robotics, Software Design, Machine Learning
Abstract: We present a cross-robot visuomotor learning framework that integrates diffusion policy–based control with 3D semantic scene representations from D³Fields to enable category-level generalization in manipulation. Its modular design supports diverse robot–camera configurations, including UR5 arms with Microsoft Azure Kinect arrays and Aloha bimanual manipulators with Intel RealSense sensors, through a low-latency control stack and intuitive teleoperation. A unified configuration layer enables seamless switching between setups for flexible data collection, training, and evaluation. In a grasp-and-lift block task, the framework achieved an 80% success rate after only 100 demonstration episodes using an UR5 robot and 90% success rate using the Aloha robot, demonstrating robust skill transfer between platforms and sensing modalities. This design paves the way for scalable real-world studies in cross-robotic generalization.

Keywords: Robotics, Machine Learning
Abstract: In self-supervised robot learning, data is acquired through active interaction with the environment, which is costly. Therefore, sample-efficient exploration is vital. To this end, intrinsic motivation (IM) methods such as learning progress (LP) have been adopted in robot learning, with variable success across tasks; whereas in machine learning, active learning (AL) is considered the go-to solution, especially for classification tasks. However, there is no systematic method for fusing both approaches for continuous regression tasks encountered in robot learning. To this end, we propose MUSEL (Model Uncertainty for Sample-Efficient Learning), a novel AL framework tailored for regression tasks in robotics, such as action-effect prediction. MUSEL introduces a novel model uncertainty metric that combines total predictive uncertainty, learning progress, and input diversity to guide experience gathering. We choose Stochastic Variational Deep Kernel Learning (SVDKL) as the base learning model and validate our approach by showing its efficacy in effect prediction tasks where a manipulator robot interacts with objects on a confined tabletop. Experiments comparing MUSEL with strong baselines show that MUSEL improves learning accuracy and sample efficiency. Overall, this study offers MUSEL as an effective online learning model applicable to any robot self-learning task where experience gathering is costly.

Keywords: Robotics, Machine Learning, Human-robot Interaction / Collaboration
Abstract: Training data diversity is an important factor in improving the performance of imitation learning. However, object placement diversity and systems that support object placement during data collection have not been sufficiently investigated. In this paper, we analyze the effectiveness of object placement diversity in imitation learning. We evaluate how different placement conditions affect task success using a simulator. Based on the results, we design optimal placement conditions and propose the object placement support system. The proposed system enabled more effective and efficient object placement than human-judged placement, and achieved comparable performance with less data.

Keywords: Automation, Plant Engineering, Network Systems
Abstract: This paper addresses the critical challenge of securing Cyber-Physical Systems (CPSs) against passive communication attacks, specifically focusing on systems modeled as Discrete Event Systems (DES). We introduce a methodology, based on the Elliptic Curve ElGamal (EC-ElGamal), to protect the confidentiality of DES-based CPS using homomorphic encryption. Our approach employs the computational intractability of the Elliptic Curve Discrete Logarithm Problem (ECDLP) to fortify supervisory control systems operating over vulnerable communication channels. The work details the EC-ElGamal protocol, from key generation to encryption and decryption, emphasizing its direct adaptation for ensuring robust data confidentiality within distributed DES architectures. A practical example demonstrating the application of this methodology in a DES context is implemented and tested on an ESP32 microcontroller.

Keywords: Robotics, Automation, Machine Learning
Abstract: Loop closure is crucial for maintaining the accuracy and consistency of visual SLAM. We propose a method to improve loop closure performance in DPV-SLAM. Our approach integrates AnyLoc, a learning-based visual place recognition technique, as a replacement for the traditional Bag of Visual Words (BoVW) loop detection method. In contrast to BoVW, which relies on handcrafted features, AnyLoc utilizes deep feature representations, enabling more robust image retrieval across diverse viewpoints and lighting conditions. Furthermore, we propose an adaptive mechanism that dynamically adjusts similarity threshold based on environmental conditions, removing the need for manual tuning. Experiments on both indoor and outdoor datasets demonstrate that our method significantly outperforms the original DPV-SLAM in terms of loop closure accuracy and robustness. The proposed method offers a practical and scalable solution for enhancing loop closure performance in modern SLAM systems.

Keywords: Automation, Hardware Design
Abstract: In aircraft manufacturing, wire harness bundling is performed manually, requiring significant time and labor. Therefore, this study presents a new design for bundling tools aimed at improving the efficiency of the bundling process. A belt-like string used for bundling is modeled using a discrete model, and constraints based on the knotting method are formulated. Under such constraints, a path of the string to knot it is optimized. From the obtained path, a mold capable of actual bundling was fabricated, and the resulting knot formed using this mold was consistent with the tying methods required for wire harness bundling.

Keywords: Software Design, Integration Platforms, Robotics
Abstract: We are developing a Cyber-Physical System (CPS) for earthwork sites, called ROS2-TMS for Construction, utilizing OPERA, an autonomous construction platform under development by the Public Works Research Institute. In this paper, as a case study of the system’s field implementation, we automated the loading of cohesive soil into a hopper during soil improvement work using a hydraulic excavator. Experimental results demonstrated that the system was able to dynamically determine excavation positions and achieve continuous operation for more than one hour. Furthermore, we conducted an additional slope-collapsing experiment using a hydraulic excavator, which demonstrated the applicability of the proposed system to diverse earthwork tasks.

Keywords: Robotics, Hardware Design
Abstract: Inflatable robots are a class of soft robots that are suitable for human robot interaction, but their control complexity and safety remain challenging. We propose a soft inflatable robot in which the exterior, continuously supplied with air by a blower, serves as the skin, and a pouch muscle embedded within the skin and actuated by air pressure serves as the muscle. This system, termed the ``musculo-skin system," forms the basis for developing an inflatable robotic arm with six degrees of freedom, including the shoulder, elbow, and four-finger gripper. The angles of the shoulder and elbow joints were measured, and the angles and motion speeds of the joints were evaluated using data from the time series of each joint. To evaluate the fingers, a Rock-Paper-Scissors gesture was performed, demonstrating the ability to execute complex finger movements without interference. The experimental results confirmed that the finger movements were smooth and capable of performing natural gestures. These results will contribute to enabling more complex motions in soft inflatable robots.

Keywords: Robotics, Mechatronics Systems, Hardware Design
Abstract: Pneumatic actuators exhibit non-linear characteristics in their pressure response to input to electromagnetic valves. This is because the discharging speed of compressed air supplied to the actuator is slower than the charging speed. The objective of this study is to develop a rapid pressure exhaust mechanism that improves the driving speed of pneumatic artificial muscles and mitigates non-linear characteristics. The proposed exhaust mechanism consists primarily of a Bowles valve and a check valve, driven by the pressure passing through them. The exhaust valve is installed at the tip of the artificial muscle, enabling pressure discharge to the atmosphere at the position closest to the artificial muscle. In experiments, the pressure response was compared when a rectangular wave voltage signal was input to the solenoid valve, verifying the validity of the proposed method. The results confirmed that the artificial muscle equipped with the developed exhaust valve exhibited a slight decrease in charging speed but a significant improvement in exhaust speed compared to the conventional method. Further optimisation of the exhaust valve design is expected to mitigate nonlinear characteristics in the future.

Keywords: Assistive Robotics, Rehabilitation Systems, Human-robot Interaction / Collaboration
Abstract: As the global population continues to age, the number of individuals with Parkinson’s disease (PD) is projected to increase significantly. Freezing of Gait (FoG), a common motor symptom in PD, affects more than half of patients and becomes more prevalent with disease progression. Suppression of FoG is essential for improving quality of life (QoL), and various assistive strategies have been proposed. This study evaluates the effectiveness of unilateral versus bilateral assistance using the Unplugged Suit for PD Patients (UPS-PD), a wearable gait assistive device powered by pneumatic artificial muscles. While prior UPS-PD studies examined only unilateral assist, bilateral assist has not been comparatively investigated. A slalom walking experiment was conducted as a single-case study with one PD participant under both assist conditions. The results indicated that bilateral assist produced more favorable gait patterns, including reduced gait asymmetry, increased swing duration, lower %DLS, and smaller variability, compared with unilateral assist. These improvements were supported by medium-to-large effect sizes in key parameters. Although generalization is limited due to the single-case design, these findings provide preliminary insights into assistive strategies for FoG suppression. Future studies involving larger and more diverse cohorts are required to validate these results.

Keywords: Human Factors, Human-robot Interaction / Collaboration, Assistive Robotics
Abstract: Postural sway and muscle co--contraction are known to increase during upright standing with aging. A previous study demonstrated that grasping a helium--filled balloon stabilized postural. In this study, we measured the center of pressure (COP) and electromyography (EMG) of the trunk and ankle muscles in 17 healthy young adults while they grasped a balloon during standing and analyzed sway and muscle activity characteristics. The results indicated that the sway--reducing effect was observed not only with upward force sensation but also with downward force sensation. These findings suggest that sensory feedback processed at the spinal cord level primarily modulates trunk muscle activity, enhancing postural control and thereby improving stability.

Keywords: Robotics, Machine Learning
Abstract: Reinforcement learning (RL) for legged robots is advancing locomotion, demonstrating its ability to adapt to new and challenging terrain. Traditionally, these RL locomotion frameworks are position-based, making the policy less adaptable to terrain types and requiring state estimation techniques in the observation space, i.e., linear velocity. Moreover, these RL frameworks often use small, lightweight quadrupeds that are limited in their viability for high-complexity tasks due to hardware constraints. This work explores an RL torque control framework for heavyweight high-torque quadrupeds. The RL framework in this paper can traverse rough terrain and effectively track a desired linear velocity without requiring knowledge of the agent's current velocity. Using Nvidia's Isaac Sim and Isaac Lab, simulation results of the RL torque control policy are shown on the Unitree B1 quadruped, achieving speeds of 3.5m/s and 1.5rad/s. In addition, the quadruped can walk up and down stairs without the aid of an exteroceptive sensor.

Keywords: Control Technologies, Mechatronics Systems, Robotics
Abstract: Inflatable habitat modules are good options for space exploration because they are light and easy to pack. However, controlling their shape and movement is difficult due to their soft materials and complex behavior. This work presents a rigid analog platform designed to support the control of such modules by implementing a second-order inverse-Jacobian control combined with null-space optimization. This control strategy allows the robot to maintain a shape leaning toward circularity while rolling in a controlled direction, improving rolling speed compared to traditional sequential control methods without compromising structural integrity. Simulations show significant improvements in speed and demonstrate the importance of maintaining the near-circular shape for enhanced performance. This approach provides a foundation for more advanced control schemes applicable to soft robotic systems.

Keywords: Robotics, Hardware Design
Abstract: Soft locomotion robots, leveraging compliant materials, adapt well to complex terrains, yet research evaluating variations in body configuration and achieving multiple locomotion modes within a single system remains limited. We present a reconfigurable soft modular robot combining urethane chip foam flexibility with modular versatility. Wire-driven actuation enables controllable deformation, allowing even a single module to move via passive body dynamics. The modular design supports rapid reconfiguration into serial connections, as well as lateral connections in either midpoint-connected (H-shape) or end-connected (V-shape) layouts. Experiments under simplified CPG-inspired oscillatory control assessed locomotion performance and terrain adaptability. The V-shape reached 34.70 mm/s, while the serial configuration with bottom-contact excelled in climbing 10 mm steps and traversing artificial grass. Adjusting actuation intervals improved deformation retention and propulsion. These results show that combining flexibility and reconfigurability enables diverse locomotion modes and robust environmental adaptability.

Keywords: Assistive Robotics, Rehabilitation Systems
Abstract: Carrying a unilateral load shifts the body’s center of mass (CoM) laterally, making foot placement adjustments critical for maintaining frontal-plane stability. The margin of stability (MoS), which quantifies the distance between the extrapolated CoM and base of support, provides a dynamic measure for assessing such stability control strategies. Although previous studies have examined segmental effects of asymmetric load carriage, they are limited in capturing whole-body dynamic stability. In this study, the impact of unilateral load carrying on dynamic stability was analyzed using MoS. Furthermore, the respective contributions of foot placement adjustments and trunk movement were explored. The findings revealed that, under moderate load conditions, MoS was primarily maintained through adjustments in foot placement. Conversely, under high load conditions, significant inter-individual variability was observed, with strategies ranging from foot placement adjustments to repositioning the CoM via trunk movement.Additionally, variations in trunk tilt direction were associated with distinct strategies aimed at reducing shoulder joint load. These findings provide new insights into the movement mechanisms underlying stability maintenance during unilateral luggage carrying and may lead to the development of assistive devices tailored to individual physical characteristics.

Keywords: Rehabilitation Systems, Machine Learning, Human-robot Interaction / Collaboration
Abstract: Electromyogram (EMG)-based motion classification using machine learning has been widely employed in applications such as prosthesis control. While previous studies have explored generating synthetic patterns of combined motions to reduce training data requirements, these methods assume that combined motions can be represented as linear combinations of basic motions. However, this assumption often fails due to complex neuromuscular phenomena such as muscle co-contraction, resulting in low-fidelity synthetic signals and degraded classification performance. To address this limitation, we propose a novel method that learns to synthesize combined motion patterns in a structured latent space. Specifically, we employ a variational autoencoder (VAE) to encode EMG signals into a low-dimensional representation and introduce a mix-consistency loss that structures the latent space such that combined motions are embedded between their constituent basic motions. Synthetic patterns are then generated within this structured latent space and used to train classifiers for recognizing unseen combined motions. We validated our approach through upper-limb motion classification experiments with eight healthy participants. The results demonstrate that our method outperforms input-space synthesis approaches, achieving approximately 30% improvement in accuracy.

Keywords: Machine Learning, Medical Devices, Human Factors
Abstract: Knee osteoarthritis (KOA) is a serious disease affecting one-third of the world's population, and mechanical loading on the knee joint has been identified as a progressive factor. However, conventional knee contact force measurements are limited to laboratory environments, making daily preventive monitoring difficult. This fundamental study aims to establish the optimal deep learning approach for three-dimensional knee contact force estimation using wearable IMU sensors and to validate the technical feasibility for practical KOA prevention systems. Methodologically, four deep learning models were systematically compared: 1D-CNN, LSTM, RNN, and Transformer. These models were trained to map 45-channel IMU time-series data comprising three-axis acceleration, angular velocity, and magnetic field from five anatomical locations to six-dimensional knee contact forces through 7-fold cross-validation. The 1D-CNN achieved the optimal maximum error of 24.01 %, while LSTM demonstrated the most stable average error of 13.93 %. These results showed superior performance compared to existing IMU-based methods, realizing three-dimensional knee joint force estimation within the practical accuracy range of wearable healthcare devices. This fundamental study successfully established the technical feasibility of translating laboratory-based knee joint load assessment to daily environments and presented clear guidelines for developing non-invasive KOA prevention systems.

Keywords: Rehabilitation Systems, Medical Training
Abstract: Observational gait analysis is commonly used in clinical settings to assess gait dysfunction and to make treatment plans. Gait Assessment and Intervention Tool (G.A.I.T.) is one of the most useful scales for observational gait analysis. However, observational gait analysis such as G.A.I.T. requires experienced clinical skills and adequate time to score. This study proposes machine learning-based prediction method of the G.A.I.T. score on individuals with stroke by pose estimation from a single RGB camera and clinical assessment data obtained in conventional rehabilitation. Twenty-five individuals with subacute stroke participated in this study. The participants were captured by single RGB camera for self-selected speed gait on the sagittal plane. In total, 40 features from conventional clinical assessment and gait parameters by single RGB camera-based pose estimation were used as input data. Five different machine learning regression models predicted the overall score of G.A.I.T. related to lower limb and trunk motions in the sagittal plane. The predicted score was compared to the actual score evaluated by an experienced physical therapist. Model performance was assessed by root mean squared error (RMSE) and coefficient of determination. The results showed that CatBoost with Boruta achieved the lowest RMSE and the highest R2 among the five models for predicting the overall score of the G.A.I.T. related to movement on the sagittal plane. This study reveals that the proposed prediction method using clinically available RGB camera gait video and clinical assessment data has the potential to predict the G.A.I.T. score on individuals with stroke.

Keywords: Human-robot Interaction / Collaboration, Robotics, Integration Platforms
Abstract: In recent years, emotion expression systems have attracted growing attention in the field of Human-Drone Interaction. However, most of these systems adopt discrete expression models, which are insufficient for capturing the ambiguity and transience of emotions. On the other hand, approaches based on continuous emotion models often rely on realistic and high-degree-of-freedom facial expressions, which are difficult to implement on non-humanoid and weight-constrained platforms such as aerial robots. To address this challenge, we propose a visual interface that continuously expresses emotion only through the movement of the eyes. In designing eyes expressions, we consider the non-biological appearance of aerial robots and emphasize friendliness and universality through an abstract visual style, inspired by the visual characteristics of emoji. Furthermore, we adopt Valence-Arousal-Dominance(VAD) model, a continuous emotional space comprising three psychological dimensions to guide the expressive behavior. Based on this model, we designed and implemented an eye display system that integratively controls visual components in accordance with the emotional state. Finally, we installed this system on a real robot, and verified its effectiveness in emotion recognition and the appropriateness of its expression through evaluation experiments. The results confirmed its effectiveness, particularly in conveying high-arousal emotional states.

Keywords: Mechatronics Systems, Renewable and sustainable energy
Abstract: Sustainable cooling is an increasingly important part of meeting the cooling demand for thermal comfort in buildings. Adsorption cooling facade systems (ACFS) are one approach to address this by utilizing solar energy on the buildings facade for sustainable indoor cooling. The ACFS’ cooling potential is provided by a facade-integrated adsorption-desorption cycle, driven by solar irradiation. The performance of the hybrid nonlinear system highly depends on ambient conditions. To efficiently operate the system, an operational strategy is necessary. This paper proposes a model-based operational strategy using optimal control. The evaluation shows a performance improvement by up to 45 % for the temperature peak overshoot above the target value compared to the rule-based operation.

Keywords: Mechatronics Systems, Automation
Abstract: To improve motor efficiency, aligned winding is crucial as it increases the slot fill factor through orderly wire placement. However, defects such as wire bulging and overlapping can arise from variations in winding speed, tension, and wire diameter, often requiring trial-and-error adjustments. While finite element methods have been used to analyze local deformation, they are computationally expensive for simulating the full winding process. We propose a mass-spring model for efficient wire simulation and extend it to three dimensions by considering wire-to-wire contact. To further reduce simulation time, we reduce the number of particles in low-impact regions and enable full 3D simulation. This method allows accurate prediction of bulging and overlapping, contributing to both winding quality and production efficiency in motor manufacturing.

Keywords: Mechatronics Systems, Hardware Design, Robotics
Abstract: Lightweight and high-strength synthetic fiber ropes are widely used as components in tendon-driven manipulators. However, measuring their tension involves structural and electrical complexity, which makes it difficult to take measurements in the middle of the rope. Through a comparison of three commercially available conductive ropes, we confirmed that the organic conductive fiber Thunderon exhibits high linearity between rope tension and electrical resistance. However, Thunderon alone has a low breaking strength, making it insufficient for application in a tendon-driven manipulator that requires high tension. To improve the breaking strength, we prototyped a composite rope using the high-strength fiber Zylon as the jacket. Tensile tests showed that the breaking strength was improved by more than 25 times and that a stable electrical response was obtained even under high tension. By implementing this rope in a coupled tendon-driven manipulator, we demonstrated that tension changes along the rope could be detected through resistance changes. This method is expected to be a new approach that enables tension measurement in coupled tendon-driven manipulators.

Keywords: Mechatronics Systems, Automation, Control Technologies
Abstract: The increasing digitization of manufacturing systems demands new strategies for quality assurance during the operational phase of computer numerically control (CNC) machines.This paper presents a novel approach to online-quality assurance using operational-parallel real-time simulation and point cloud comparison techniques. During machining, a real-time simulation generates a virtual representation of the manufactured workpiece, reflecting the actual position values and process conditions. This simulated point cloud is then compared directly to the nominal CAD model to detect geometric deviations early in the production process. Unlike conventional approaches relying on post-process optical measurements, this method enables in-process evaluation without interrupting operations. It supports fast, non-invasive quality assessment and facilitates adaptive process control. The work presented in this study presents an evaluation of different point cloud comparison methods for an online-quality assurance framework.

Keywords: Virtual / Augmented / Mixed reality, Human Factors
Abstract: With conversational artificial intelligence (AI) becoming increasingly present in our daily lives, some people have begun to view it as a friend, rather than merely a tool. Among young people in particular, the use of conversational AIs as confidants is becoming more widespread, signaling the emergence of a new form of interaction between humans and AI.However, current conversational AIs are limited to text-based communication and lack the nonverbal cues found in human communication, such as facial expressions, intonation, and gestures. This lack of nonverbal cues raises concerns that the user might miss the nuanced richness of meaning that is intended to be conveyed. To expand the expressive capabilities of future conversational AIs, we propose an avatar that conveys facial expressions, which play an important role in conveying emotions. Based on Plutchik's emotion model, the avatar can display eight primary facial expressions and composite expressions by combining them. We experimentally investigated how participants perceive the expressions of the proposed avatar. This experiment revealed that anger, fear, disgust, and trust are difficult to recognize. Additionally, the results suggest that, even when male and female avatars express the same emotions, they may be perceived slightly differently.

Keywords: Assistive Robotics, Human-robot Interaction / Collaboration, Machine Learning
Abstract: This study proposes a Transformer-based approach with cross‑attention for predicting human facial movements in face‑related robotic control tasks and integrating these predictions into an imitation learning framework. A dataset of human facial videos was constructed, and landmarks were extracted using the MediaPipe framework. Three prediction methods were compared, and the cross‑attention model achieved the best performance in both landmark localization accuracy and image quality. In imitation learning experiments, facial motion trajectories sampled from real human data trajectories were used, and the success rate increased from 42% to 60% and ultimately to 74% when predicted landmarks were incorporated. Additionally, varying the prediction horizon affected task completion time, with the 2‑frame horizon achieving the fastest completion. These results demonstrate that incorporating predicted facial motion can significantly enhance robotic control performance in dynamic human‑robot interaction scenarios.

Keywords: Telecommunication Systems, Machine Learning, Human-robot Interaction / Collaboration
Abstract: To enable natural non-verbal communication during immersive avatar control, it is essential to accurately reproduce an operator's intended facial expressions on an android avatar. This study proposes a facial synchronization system that converts facial parameters captured from head-mounted displays (HMDs), into the corresponding expressions on an android avatar. In conventional systems, actuator commands controlling facial movements are typically mapped linearly to facial parameters. However, the correspondence between android facial actions and facial parameters cannot be clearly defined. We developed a motor command mapping model (the “developer model”) that generates the developer's intended avatar expressions based on facial parameter data and actuator commands collected through repeated facial mimicry. We then proposed and evaluated a “personalized model,” which adapts the developer model to individual operators to better replicate their intended expressions. In a participant experiment comparing three models (conventional, developer, and personalized), no significant differences were observed. However, the developer model—despite lacking personalization—received evaluations comparable to the conventional model, supporting the effectiveness of the proposed approach. In contrast, the personalized model received mixed evaluations, with many participants rating it lower than the other two models, indicating a need to improve the personalization process. This study offers insights for developing more effective facial synchronization systems for immersive android avatars.

Keywords: Human-robot Interaction / Collaboration, Assistive Robotics, Robotics
Abstract: While the conventional stroking motions using robot arms in physical human-robot interaction were planned from the pre-recognized shape of the target, it's not practical because the movements of the target body during stroking are not considered. In this paper, we propose a stroking motion generation method using a mobile manipulator and an RGB-D camera to follow changes in the target human's posture. Specifically, we first extract the target subject with image recognition and segmentation, and point cloud processing. Next, we generate the target part's 3D model and a motion trajectory on the model. Finally, we repeatedly update the trajectory by following changes in the target part's posture using an Iterative Closest Point (ICP) algorithm. For evaluation, the proposed method was implemented on a mobile manipulator HSR. Then, stroking motions were generated on a human-shaped robot's back with various movements, and the alignment success score and the alignment error were measured. From the results, we confirmed that the motion generation was highly successful and occlusion by the manipulator's arm affected the alignment.

Keywords: Medical Devices, Mechatronics Systems, Robotics
Abstract: This paper proposes a urinary catheter insertion assistance mechanism aimed at reducing pain and the risk of injury during insertion, as well as an agarose–PEG gel urethra model simulator for its performance evaluation. Conventional urinary catheterization has the drawback that friction against the urethral wall and impingement at curved sections can cause severe pain and tissue damage to patients. To address this, we focused on a soft growing actuator that can advance through a lumen without sliding its outer surface, and we developed a urinary catheter insertion assistance mechanism utilizing this actuator. In addition, we newly developed a gel-based urethra model simulator, made of agarose and PEG, capable of reproducing urethral length, diameter, curvature geometry, and Young’s modulus, and used it for performance evaluation of the proposed mechanism. Fundamental property tests confirmed that gel containing 2% agarose and 20% PEG achieved a Young’s modulus (~125 kPa) equivalent to that of the urethra and exhibited high toughness. In insertion experiments using the urethra model simulator, the proposed mechanism demonstrated stable extension and catheter traction even in models with curvature or straight section stenosis. These results indicate that a urinary catheter insertion assistance mechanism employing a soft growing actuator has the potential to effectively reduce friction and alleviate pain inside the urethra.

Keywords: Medical Devices, Assistive Robotics, Medical Training
Abstract: This study presents a novel ECG monitoring device named the Fabric Actuator with Embedded Electrode (FAEE) that seamlessly integrates skin condition regulation to improve biosignal acquisition, particularly under dry skin conditions, capable of gentle interaction with the human body. The FAEE incorporates a fabric actuator that actively modulates temperature and humidity, alongside textile-based electrodes for biosignal acquisition, enabling reliable heart rate monitoring. The actuator is constructed with fluidic channels embedded in a fabric engineered to be both waterproof and moisture-permeable, allowing for controlled water circulation and vapor release. By regulating skin moisture levels, the actuator enhances electrode-skin contact, leading to stable and high-quality ECG signal acquisition. The soft, adaptive characteristics of the developed ECG monitoring system offer not only functional monitoring capabilities but also psychological comfort.

Keywords: Machine Learning, Decision-making systems, Medical Devices
Abstract: Cancer therapy–related cardiac dysfunction (CTRCD) is a serious complication, posing significant risks to cancer survivors' cardiovascular health. While early detection is crucial, current imaging-based methods are expensive and impractical for routine screening. This study proposes a deep learning framework that combines convolutional neural networks (CNNs) and graph convolutional networks (GCNs) to analyze 12-lead electrocardiography (ECG) data for CTRCD detection. By representing ECG leads as graph nodes and modeling their anatomical and physiological relationships through different adjacency matrices, our approach captures inter-lead dependencies overlooked by conventional methods.

Experimental results demonstrate that the physiological function-based graph structure outperforms the conventional CNN approach, particularly in sensitivity and F1-score. Interpretability analysis reveals distinct lead-specific patterns, enhancing clinical understanding.

Keywords: Robotics, Software Design, Mechatronics Systems
Abstract: Finite element modeling (FEM) provides a powerful and flexible approach for accurately simulating soft robots and their dynamics. This work presents a simulation-based framework for identifying the mechanical properties of a 3D-printed, cable-driven soft robotic neck made of thermoplastic polyurethane (TPU) with a Shore hardness of 82A. A Bayesian optimization method is employed during FEM simulations in the open-source SOFA framework to estimate material parameters by minimizing the discrepancy between experimental data and simulation results. Cyclic traction tests were conducted on a custom test bench to obtain force and displacement data from one of the three actuation cables of the neck. Based on these observations, a hyperelastic constitutive model was selected, and its parameters optimized. The simulated response remained consistent with experimental measurements, indicating that the model captures the essential mechanical behavior of the system. This approach opens a practical alternative to direct material characterization and lays the foundation for future extensions involving viscoplastic models adaptive force-based control.

Keywords: Robotics, Automation, Control Technologies
Abstract: In this paper, we propose a novel DBSCAN method for 3D object segmentation that considers point cloud density based on lidar range and scanning order of laser beams. Point clouds obtained by a lidar are dense at short ranges and sparse at long ranges. Furthermore, point clouds obtained by the lidar are organized in the scanning order of each beam. However, conventional DBSCAN has issues that its parameters are fixed and cannot adapt to density variation of point clouds, and that the computational costs for neighborhood search are high, resulting in significant processing time. Therefore, in the proposed method, parameters are adjusted according to lidar ranges to accommodate density differences. Furthermore, executing neighborhood search within only a specific beam scanning area reduces computational costs and shortens processing time. We conducted experiments in multiple environments to verify the effectiveness of the proposed method. By automatically determining parameters based on the point cloud density, the proposed method demonstrated that both nearby and distant objects can be correctly segmented. Additionally, our method demonstrated that processing time can be reduced by executing neighborhood search within only a specific scanning area of the beams. As described above, the proposed method achieves adaptive and high-speed object segmentation by considering both the differences in point cloud density due to lidar ranges and the beam scanning order.

Keywords: Robotics, Machine Learning, Intelligent Transportation Systems
Abstract: As the use of Unmanned Aerial Vehicles (UAVs) in various tasks within human-inhabited environments becomes increasingly common, critical aspects such as emergency landing need to be addressed. The use of deep learning has become widely adopted to provide context-sensitive solutions, where semantic segmentation has shown promising results. Therefore, this paper presents a comparative study that aims at evaluating several candidate semantic segmentation algorithms, offering a guide for an appropriate selection of the segmentation module in a UAV safe landing task in unstructured urban environments. More specifically, a comparison was made between three prominent segmentation models, U-Net, SegFormer, and MANet, using the Semantic Drone Dataset. First, the models were evaluated using the original 24 classes of the dataset. SegFormer performed slightly better than the other algorithms tested. In a second, more critical experiment, the classes were grouped into six risk levels based on the Specific Operations Risk Assessment (SORA) framework. To address class imbalance and prioritize high-risk categories, a weighted Cross-Entropy loss was employed, assigning higher penalties to misclassifications in critical risk levels. In this setup, MANet achieved the best results, showing its ability to adapt to risk-based classification and capture important features. Later, the three models were optimized to prove that even complex architectures can be enhanced for real-time inference. The optimization included converting the model to TensorRT format and using FP16 precision, which reduced the models' size by at least 30%. Finally, the optimized U-Net, the largest model, was tested on a NVIDIA Jetson Orin Nano platform achieving real-time inference. This shows that heavy models can run on embedded hardware like the Jetson Orin Nano, making them viable for safe, autonomous UAV applications.

Keywords: Environment / Ecological Systems, Machine Learning, Decision-making systems
Abstract: This work proposes an alternative solution to address the problem of tree canopy classification, given a digital elevation model (DEM) of the study area obtained using a Light Detection and Ranging (LiDAR) device. This proposal presents a comprehensive methodology that enables the feasible application of machine learning (ML) models for tree canopy classification. Our approach allows to obtain a composite DEM of the tree canopy, which will be characterized using a tree cover selection and extraction strategy based on Euclidean distance and feature point clustering. Subsequently, statistical and geometric measures of the point clouds are calculated for each segmented canopy of the DEM in order to prepare a dataset, which includes some tree canopy features represented by column vectors of 77 predictors. Finally, the generated dataset was trained and validated using the MatLab's classification learner, obtaining considerable performance results in canopy classification, highlighting the improvement in classification performance of models based on Naive Bayes, Neural Networks, and Support Vector Machines. Furthermore, beyond its technical scope, this research has a direct social and environmental impact by promoting urban resilience, biodiversity conservation, and strengthening sustainable development policies. The proposed methodology contributes to global agendas such as the United Nations Sustainable Development Goals (SDGs) and the Mexican National Strategic Programs (PRONACES), positioning remote sensing and artificial intelligence as tools for climate action and social well-being.

Keywords: Robotics, Control Technologies
Abstract: We present a novel approach that integrates photometric Image-Based Virtual Visual Servoing (IBVVS) with Gaussian Splatting (GS), a recent and efficient 3D representation for photo-realistic view synthesis. In our framework, the servoing process is performed entirely in simulation using a GS model trained on a sparse set of images from a scene wheras unseen images serve as target views for photometric IBVVS. At each iteration a rendered image from the GS model simulate the camera’s current view, and the pixel-wise intensity error between the rendered and target images is used to compute control commands for camera pose optimization. This framework removes the need for externally acquired explicit 3D geometry or precomputed dense depth maps from traditional sensors, since depth information is implicitly obtained from the GS representation and used directly in the control loop. The method enables virtual servoing toward novel views that were not captured during training. Experimental results demonstrate accurate and smooth convergence, highlighting the potential of learned view synthesis for 3D camera tracking and visual servoing applications.

Keywords: Robotics, Network Systems, Machine Learning
Abstract: This study proposes a hierarchical cooperative control architecture for multi-robot exploration, centered on a dynamic branching–integration mechanism that adapts swarm structures to environmental conditions. The architecture integrates a System Agent for global coordination and a Swarm Agent for local action decisions, enabling adaptive and scalable cooperation.

Simulations in multiple environments show that the branching–integration mechanism alone surpasses the baseline in all cases, with gains in complex settings. Learning-based optimization of branching–integration conditions further improves early-stage exploration speed and robustness.

These results indicate that dynamic swarm reconfiguration is the primary driver of exploration efficiency, while learning effectively enhances its benefits. Future work will investigate other reinforcement learning algorithms and exploration strategies, evaluate performance under dynamic or communication-constrained conditions, and conduct field tests with real robots.

Keywords: Machine Learning, Mechatronics Systems, Automation
Abstract: With the frequency of wildfires increasing, specifically across the Western United States, there exists a need for the reduction in damage to both the community and environment. Many urbanized areas and communities have been significantly affected or demolished by the rapid and devastating spread of recent wildfires. As the impact of global warming increases and the weather patterns begin to vary, the ability to contain a wildfire quickly begins to pose a challenge. While current solutions exist on the market, they are simply not designed for the increasingly faster spread because of ever-changing weather patterns. Additionally, the need for firefighters to have specialized experience has caused a rift between the number of those who can contain fires and the number of fires to be contained. The purpose of this project is the development of a machine learning wildfire UAV swam algorithm to allow the rapid and real time update of containment line calculations/visualizations to allow first-responders to effectively and quickly contain wildfires. By measuring major heat signatures, the UAV system can identify wildfires, and through TensorFlow analysis of wind patterns and environmental markers, make feedback decisions of where to search for spot fires. The combination of wildfires and the spot fires that may arise, result in a containment line visualization with coordinates for first responders to utilize. The goal of the IPCA is a functional autonomous system that can provide the prediction and aid for first responders to have another line of defense against the spread of wildfires into residential and urbanized areas.

Keywords: Robotics, Network Systems, Intelligent Transportation Systems
Abstract: To detect obstacles in narrow spaces, such as under-floor area, herein, we proposed a method that estimates the location of obstacles using a distributed model of swarm robots. First, we constructed a swarm robot model, in which a robot moves by jumping at regular intervals in a field enclosed by walls. We confirmed that, when no obstacles were placed in the field, the spread of the swarm robots followed a Gaussian distribution over a certain period. We then assumed that, even when obstacles were present, the distribution of the robots in the field would follow the Gaussian distribution except in the neighborhood of obstacles. Under this assumption, we proposed a method to clearly estimate the positions of obstacles by taking the difference between the approximated Gaussian distribution based on an average gross distribution of robots over a certain period and the actual average gross distribution of robots in each subdivided area of the field.

Keywords: Intelligent Transportation Systems, Decision-making systems, Large-scale Infrastructure Systems
Abstract: Intermodal journey planning remains a challenge in intelligent transportation systems, particularly when accounting for heterogeneous passenger preferences and the integration into smart cities. Traditional planning approaches often fail to capture dynamic traffic conditions and the passenger-centric view required for future transportation systems. This study proposes a Multi-Objective Multi-Agent Reinforcement Learning (MOMARL) framework for individual intermodal journey planning across multiple modes. Two microscopic traffic model were developed in Simulation of Urban Mobility (SUMO), creating simulation environments in which passengers plan their journeys to arrive on time at transportation hubs. One simpler model for the verification of the framework and a calibrated model reflecting the dynamics of a real city. The transportation networks were modeled as multilayered graphs. Since each passenger has different preferences and access to transport modes, their individual cost-minimal paths are formulated as a multi-objective optimization (MOO) problem. From this, the scalarized reward signals used in the MOMARL framework are derived. Simulation results show that the proposed approach enables agents to generate feasible intermodal routes in a microscopic environment, demonstrating the use of MOMARL for passenger-centric coordination in multimodal transport systems. Application to the calibrated model of Ingolstadt posed challenges regarding simulation complexity, highlighting the need to expand research in methods that allow the systematic reduction of model fidelity and granularity while retaining realistic dynamics.

Keywords: Robotics, Integration Platforms, Mechatronics Systems
Abstract: This paper presents the design, fabrication, and evaluation of a textile-based strain sensor integrated into a soft robotic joint for real-time monitoring of bending angles. The proposed sensor consists of a conductive fabric encapsulated in a flexible silicone substrate, created using a custom co-casting process that enables seamless integration without compromising structural compliance. Electromechanical characterization under tensile and compressive loading demonstrated a strong piezoresistive response, with a gauge factor (GF) of 77.76 under tensile strain and 9.24 under compression, within working ranges of 6–12% and 3–14%, respectively. The sensor was embedded into a soft joint with asymmetric geometry to assess directional sensitivity and real-time performance. Experimental results confirmed reliable bidirectional strain detection and an over 500% increase in resistance under tensile bending. A second-order polynomial model accurately mapped the sensor’s resistance changes to joint bending angles (10°–25°), achieving 𝑅2 values above 0.67 in both tensile and compressive configurations. These findings validate the sensor’s potential for real-time posture estimation in wearable and soft robotic systems.

Keywords: Robotics, Control Technologies, Automation
Abstract: This paper proposes an end-effector position and attitude estimation method for hydraulic excavators using total stations and potentiometers. The total station is an external sensor for measuring positions in a ground-fixed coordinate, but its sampling interval is too long for a position controller. The potentiometer is a joint angle sensor whose sampling interval suits the controller. The proposed method is an extended Kalman filter that simultaneously estimates position and attitude. It combines measurements from total stations and potentiometers with different sampling intervals to estimate accurate end-effector position and attitude with a short sampling interval. The proposed method's validity and applicability to the end-effector position control are verified using a real-time simulator of a hydraulic excavator.

Keywords: Robotics, Control Technologies, Automation
Abstract: This paper proposes a method for generating end-effector trajectories for automatic excavators to prevent tipping over and avoid collisions with obstacles. The criterion for determining whether the excavator is at risk of tipping over is derived using the Zero-Moment Point (ZMP), which represents the center of reaction forces. A simplified model of the excavator is used in this derivation to avoid complexity. Additionally, the end-effector trajectories are automatically generated by using an algorithm based on the penalty function method, which incorporates constraints to prevent both tipping over and collisions with nearby obstacles. The automated excavator followed the generated trajectories in a simulator, and it was confirmed that tipping over was prevented and collisions with obstacles were avoided.

Keywords: Robotics, Human-robot Interaction / Collaboration, Software Design
Abstract: This paper presents two types of extended diffeomorphism designs to compensate for spatial placement differences between robot workspaces. Teleoperation of multiple robots is attracting attention to expand the utilization of the robot embodiment. Real-time reproduction of robot motion would facilitate the efficient execution of similar tasks by multiple robots. A challenge in the motion reproduction is compensating for the spatial arrangement errors of target keypoints in robot workspaces. This paper proposes a methodology for smooth mappings that transform primary robot poses into follower robot poses based on the predefined key points in each workspace. Through a picking task experiment using a dual-arm UR5 robot, this study demonstrates that the proposed mapping generation method can balance lower mapping errors for precise operation and lower mapping gradients for smooth replicated movement.

Keywords: Human-robot Interaction / Collaboration, Human Factors, Hardware Design
Abstract: Wearable haptic feedback has become an essential component in human-robot interaction, including teleoperation, prosthetics, and Supernumerary Robotic Limbs (SuperLimbs). However, most existing systems primarily convey contact or force information, and few address proprioception, the sense of body position and movement, which plays a key role in human motor control. In this paper, we propose a proprioception-inspired haptic feedback method using a wearable device that informs users of the movement state of a robot's end-effector. We implement two feedback strategies: Movement Haptic Feedback (MHF) and Error Haptic Feedback (EHF). Two user studies were conducted under disturbed visual conditions: one in a teleoperation scenario and another using a SuperLimb scenario. Results showed that both haptic feedback approaches improved task performance and reduced subjective workload compared to no-feedback conditions. These findings demonstrate the value of proprioception-inspired haptic design in expanding body awareness in human-robot sensory augmentation.

Keywords: Human-robot Interaction / Collaboration, Robotics, Control Technologies
Abstract: Training humans to control dynamic systems such as prosthetic limbs or teleoperated robots typically requires extensive practice. In this study, we investigate whether passive exposure to skilled control behavior—delivered via synchronized haptic and visual playback—can accelerate visuomotor learning in a non-trivial control task. Using a simulated cart-pole environment, we compare performance between participants who received an initial session of passive visuo-haptic replay and those who directly began with active control. Our results show that passive exposure yields two clear benefits: improved initial performance and a steeper learning curve across subsequent sessions. These findings suggest that passive sensorimotor experience, even without haptic guidance during active control, can support the acquisition of motor skills necessary for dynamic control. This approach may provide a low-effort and scalable training paradigm for enhancing skill acquisition in robotic and assistive technologies.

Keywords: Network Systems, Assistive Robotics, Software Design
Abstract: Soft robotics, with its flexible and lightweight structures, offers innovative solutions in fields such as medicine, rehabilitation, and assistive technologies. The development of textile-based soft robotics has significantly reduced integration challenges in wearable technologies. This study presents a novel fog computing architecture designed to support multi-user and multi-device operations used in textile-based soft robotics systems. For patients experiencing muscle weakness in one hand, Textile-based Internet of Things (T-IoT) gloves enable simultaneous bilateral actions such as grasping, lifting, and carrying through a mirroring technique. Sensor data from a sensing T-IoT glove worn on the healthy hand are processed by the fog computing system and transmitted to an actuating T-IoT glove worn on the impaired hand, enabling coordinated movement of both hands. By integrating various concurrency and interprocess communication techniques into worker devices, the system can serve up to 10 users with a single worker and up to 26 users with three workers operating simultaneously. In this way, the proposed system provides real-time computational services for multi-user scenarios while also delivering an assistive technology for patients.

Keywords: Robotics, Control Technologies, Mechatronics Systems
Abstract: High temporal resolution sensors have become increasingly widespread in recent years, enabling detailed and accurate observation of fast physical phenomena and contributing to the understanding of their essential properties. However, a fundamental challenge remains in bridging the temporal resolution gap between such sensors and conventional industrial robots. Typical high-speed visual feedback control require all system components to operate at high speed, inherently restricting their applicability to phenomena within the robot's dynamic capabilities. As a result, valuable high-speed sensing data often cannot be fully utilized, especially when target phenomena evolve faster than the robot can respond.

This study proposes a novel motion planning method that extracts essential, task-relevant information from high-speed sensor data to bridge the timescale gap between sensors and robots. Based on the extracted information, feedforward control is executed with consideration of the robot's motion characteristics. To explore practical solutions, we developed an integrated system combining a high-speed camera and a proximity sensor module—both with existing applications in robotics—with a general-purpose industrial robot and gripper. A case study involving the grasping of a pendulum-swinging object by a low-speed robot demonstrates the effectiveness of the proposed approach in utilizing high-frequency measurements for tasks beyond the limits of conventional feedback control.

Keywords: Robotics, Machine Learning, Automation
Abstract: Research aimed at enabling robots to acquire not only repetitive and simple tasks but also complex and delicate tasks is receiving significant attention to expand the use of robots to support and replace workers. Imitation learning is one of the promising approaches to enable robots to learn complex human skills with minimal learning cost. However, since the motion data generated by human-operated robots serves as the reference motion for the robot, it is difficult for the robot to acquire fast and smooth motions that surpass human operation. In this paper, we propose a data augmentation method to generate faster and smoother motion data by adjusting the robot’s motion output from the trained model. By utilizing model predictive control (MPC) for motion adjustment, it becomes possible to balance trackability to the original motion and smoothness of the motion through the design of the evaluation functions and constraints. The robot can perform tasks efficiently and robustly by continuously learning augmented motion data that has been optimized using MPC. We demonstrate through experiments on object picking and placing task that higher-quality motion data generated in the real-world.

Keywords: Robotics, Automation, Machine Learning
Abstract: This paper proposes a method for topological mapping that guarantees orientation accuracy. Unlike pure topological maps, the orientations of arcs connecting nodes in our topological maps match those in real environments. The proposed method uses the Visual Place Recognition (VPR) method, AnyLoc-VLAD-DINOv2, to extract global descriptors from images obtained by a 360-degree camera, and then creates nodes and detects loops based on these descriptors. Furthermore, the initial orientation of each node is calculated using angular velocity obtained by an IMU. The proposed method creates two types of orientation constraints. Heading constraints are created from initial orientations, and loop constraints are created from loop detection. Subsequently, node orientations and arc lengths are corrected through constrained optimization with two types of orientation constraints. Through indoor and outdoor experiments, the proposed method enabled topological mapping with both topological consistency and orientation accuracy. Nodes of the topological maps were created adaptively based on the appearance of each location within the environments, and the node spacing varied accordingly. Furthermore, through constrained optimization with two types of orientation constraints, node loops were closed and arc orientations matched those in real environments. Even in environments containing multiple loops, the proposed method enabled topological mapping while simultaneously satisfying the constraint of each loop.

Keywords: Robotics, Control Technologies
Abstract: Autonomous mobile robots are expected to demonstrate a high degree of adaptability, enabling effective operation across diverse environments. Robust localization is a key requirement for achieving such autonomous mobility. Typically, localization methods using sensors such as cameras and LiDAR construct a map in advance from features of the driving environment. The robot then estimates its pose by matching the currently observed features to the pre-prepared map. However, environmental changes create discrepancies between the current environment and the pre-prepared map, leading to localization failure. This paper presents a map selection method for robust localization, which selects a map that reflects the current environment from multiple pre-prepared maps constructed under different environmental conditions. Although existing methods are limited to a single sensor, the proposed method can be applied to different types of sensors in a unified manner by handling sensor information at the pose information layer. To achieve this, the method utilizes the relative pose difference between the interoceptive sensor and the map, which is less susceptible to environmental changes, thereby enabling appropriate map selection under varying environmental conditions. The experiment was conducted using a robot equipped with a stereo camera in an environment with four conditions. The results showed maximum localization errors of 0.10–0.24 m and mean errors of 0.03–0.04 m, demonstrating robust localization through the selection of an appropriate map that reflects the current conditions.

Keywords: Human-robot Interaction / Collaboration, Robotics, Decision-making systems
Abstract: In physical Human-Robot Interaction (pHRI), a significant challenge lies in the unpredictable nature of human behavior, which can introduce a high level of uncertainty during an interaction with a robot. While traditional force-based control laws lack high-level reasoning, learning-based methods do not treat perception and action uniformly to decrease uncertainties about human intention. This paper presents a pHRI framework based on Active Inference (AIF) for planning and decision-making, which guides a robot to balance goal-directed behavior and the exploration of the human intention. The framework integrates a 1D-CNN based Conditional-Variational Autoencoder (CVAE) architecture for Expected Free Energy (EFE) computation and policy selection with a Cartesian impedance controller to allow the robot to adapt its motion according to the selected policy. Additionally, this paper investigates the influence of preference precision on the robot’s behavior. The pHRI experiment includes pushing, pulling and no-applied interaction scenarios. The results show that the robot naturally favors goal-directed, high-stiffness behavior when it is undisturbed, while it prefers exploration behavior for lower preference precision values and goal-directed behavior for higher preference precision values during pushing and pulling interaction scenarios. The findings of this research demonstrate that the preference precision parameter significantly influences the process of minimizing uncertainty of human behavior, enabling the robot to adaptively balance exploration and goal- directed behavior in pHRI scenarios.

Keywords: Robotics, Control Technologies, Integration Platforms
Abstract: Conventional CNC controllers, originally developed for machine tools, are increasingly used to control robotic systems. However, their decentralized axis structure and kinematic abstraction limit dynamic performance compared to dedicated robot controllers. To address this limitation, this paper presents a methodology for integrating dynamics-based feedforward control within CNC-controlled robots by reusing digital twins from virtual commissioning. The approach extends existing virtual models with inverse dynamics through standardized interfaces and deploys them directly on the CNC controller as real-time digital twins. These models generate feedforward torque signals that complement conventional feedback control, improving trajectory tracking without modifying the CNC architecture. The main contibution lies in a generalizable workflow that links simulation, commissioning, and operation within a unified framework. The method is validated experimentally on a Delta robot, demonstrating significant reductions in tracking error and showing that the approach can be readily adapted to other CNC-controlled robotic systems.

Keywords: Robotics, Control Technologies, Mechatronics Systems
Abstract: Contact task execution in unknown environments is fundamental to robotic applications, requiring three essential functions: accurate position tracking, safe contact establishment, and achievement of desired contact force. In our previous work, the Multiple Virtual Dynamics-based Force Control (MVDFC) was proposed. This method seamlessly integrates these three functions, however, several challenges remain when considering practical force control. The first challenge is that time delays in the force sensor can destabilize the control system. The second is that loss of contact with the environment during a force control task can result in acceleration of robot and collisions with the environment. These two challenges are commonly encountered in various force control methods. Here, a key feature of MVDFC is its flexibility, allowing the motion of virtual objects in each virtual dynamics to be independently designed. Therefore, by leveraging this flexibility, it is possible to overcome the two challenges without compromising the three functions. This study proposes a direction for parameter design to address the above issues, and its effectiveness is demonstrated through both simulations and experiments.

Keywords: Control Technologies, Machine Learning, Robotics
Abstract: We present a fiducial marker system tailored for underwater acoustic imaging, enabling accurate detection and recognition of multiple marker IDs in real-world Forward-Looking Sonar (FLS) images. The marker is physically designed with layered concrete–metal structure to generate strong and distinctive sonar reflections. Our marker detection and recognition pipeline is trained entirely on simulation data, yet it achieves accurate performance on real-world sonar images. By leveraging a custom FLS simulator we generate annotated training samples that closely mimic real sonar characteristics. A YOLO-based detector, trained with these simulated images, localizes markers and regresses corner keypoints. For marker identity recognition, detected regions are rectified and decoded using a grid-based binary recognition scheme. Experiments show that the model achieves a 86.6% true positive detection rate and 100% ID recognition accuracy in the fully visible patch subset of real sonar images, despite being trained solely on synthetic data. This sim-to-real framework offers a scalable solution for underwater localization and inspection in autonomous robotic systems.

Keywords: Control Technologies, Robotics, Software Design
Abstract: This paper proposes a control method for robotic hammering that can attenuate the reaction torques on the robot's joints while maintaining a sufficient impact on the environment. A key challenge in hammering tasks is this trade-off between delivering an impact and preventing damage to the manipulator. The proposed method addresses this conflict by generating the entire pre-collision trajectory, including the required initial posture, by calculating backward in time from a desired state at the moment of impact. The trajectory consists of a preparatory backswing and a striking motion, both tracked using a PD controller. To further attenuate reaction torques, a feed-forward torque pulse is applied to the final joint just before impact. After impact, a force-controlled lift-up mode is activated to raise the hammer and return the robot to its initial posture. The effectiveness of the proposed method is validated through experiments, demonstrating an attenuation in joint torque.

Keywords: Robotics, Mechatronics Systems, Hardware Design
Abstract: This paper presents a comprehensive study on the design and performance of a rotational spring mechanism for Series Elastic Actuators (SEAs), fabricated using 3D printing techniques. The spring consists of six symmetrically arranged elastic blades that connect concentric inner and outer rings, enabling bidirectional compliance and large angular deflection. By systematically varying geometric parameters, particularly blade thickness, we investigate the influence on torque-displacement behavior and overall stiffness. A lumped parameter model is developed to analyze the spring mechanics under rotational deformation, and closed-form expressions for torque response are derived. An experimental setup is developed using a DC motor to apply torque to the designed springs and assess the torque-displacement relationship. The results provide key insights into how design parameters affect spring performance and offer practical guidelines for the development of lightweight, customizable compliant elements in next-generation robotic actuator systems.

Keywords: Robotics, Mechatronics Systems, Assistive Robotics
Abstract: This study focuses on the structure and actuation principle of biological muscles and attempts to develop a novel actuator that emulates their behavior. Pneumatic artificial muscle (PAM), particularly those of the McKibben type, have been widely utilized in assistive suits due to their lightweight design, flexibility, and high-power density. However, their reliance on compressed air results in delayed response, making them unsuitable for rapid actuation. To address this limitation, we developed a combustion-driven artificial muscle named HADEC (High-Response Artificial Muscle Actuator using Dimethyl Ether Combustion), which generates impulsive contraction force by injecting and igniting a dimethyl ether (DME)– air mixture inside the artificial muscle. Since HADEC is only capable of impulsive behavior, it has not yet been able to achieve steady-state force generation. In this work, we demonstrate a system architecture in which three HADEC units are driven at high frequency with phase differences. The proposed configuration successfully achieves steady-state force generation. Experimental results show that an arm could be held within a range of 10°–20° for approximately 1.5 seconds, indicating that repeated high-frequency actuation of HADEC can emulate the behavior of biological muscles and serve as an effective approach for achieving continuous actuation.

Keywords: Robotics, Software Design, Control Technologies
Abstract: Given recent technological trends, such as the use of Digital Twins, modeling and simulation have become essential in both the design and the operation of robotic systems. However, developing large simulations that feature several aspects of a robotic system is not trivial. Co-simulation can be used to address this challenge by the hierarchical aggregation of heterogeneous simulators. The recent industry standard Functional Mock-up Interface version 3 (FMI3) adopts the concept of super-dense time to enable both continuous-time and discrete-event simulation. In this work, we extend an existing framework for robotics that supports the creation of modular co-simulation scenarios with simulated or real robots in the loop by upgrading its artifacts to support hybrid co-simulation with FMI3. With these upgraded artifacts, the extended framework accounts for both continuous-time and discrete-event simulations, providing a more comprehensive coverage of scenarios and behaviors of robotic systems. We use a UR5e robotic arm for the demonstration of the FMI3-enabled framework and provide our findings and recommendations from the experiments executed.

Keywords: Robotics, Control Technologies, Automation
Abstract: With the growing use of industrial robots in high-force operations, accurate dynamic modelling has become increasingly critical for their design and control. The authors previously proposed a novel automated method to identify joint dynamic parameters of moving industrial robots, considering frictional behavior across a large portion of the workspace. This method excites the robot using a “fast chirp” centrifugal force, allowing precise control of the excitation force. However, its accuracy is limited due to the neglect of out-of-plane z-dynamics and the error in the input force model. This paper addresses these limitations to enhance the model’s prediction accuracy. Theoretical and experimental excitation forces are compared to assess their influence on identification. The indirect z-direction vibration response is formulated accounting for the cross-couplings and analyzed to enhance the prediction accuracy of out-of-plane z-dynamics. The experimental results show that the proposed approach reduces prediction error by up to 49.8% compared to the previously identified model, in which the z-direction vibration responses were not considered.

Keywords: Decision-making systems, Renewable and sustainable energy, Large-scale Infrastructure Systems
Abstract: In this paper, we address the design of an energy management system (EMS) that incorporates a mechanism to promote the cooperation of electric vehicle (EV) users. The proposed EMS incentivizes EV users to cooperate, particularly to remain at home and make their EVs available for discharging power to the grid as reserve capacity. To this end, we model the behavioral change of EV users in response to incentives. Then, we formulate model-based incentive-driven EMS (ID-EMS) design as an optimization problem and derive its analytic solution. Finally, we validate the effectiveness of the ID-EMS through numerical experiments. The experiments leverage real-world data on daily vehicle usage patterns and survey-based insights into user behavioral changes.

Keywords: Renewable and sustainable energy, Control Technologies, Mechatronics Systems
Abstract: Thermochemical energy storage based on the reversible reaction of CaO and H2O to Ca(OH)2 is a promising solution for sustainable storage of excessive renewable energy. For efficient and safe discharge of the system (hydration of CaO), a control strategy is necessary. Starting from a nonlinear dynamic model, the system is linearized and discretized to enable the design of a Kalman filter for state estimation and a Model Predictive Controller (MPC). Simulation results demonstrate that the Kalman filter provides accurate state reconstruction and effectively filters measurement noise from the system outputs. Four different MPC objectives, targeting key temperatures and thermal power, are evaluated. All variants show good tracking performance and are suitable for real-time application. A comparison between the linear discrete-time MPC with state estimation and a nonlinear continuous-time MPC with full state information reveals no significant performance loss, while achieving a reduction in computation time.

Keywords: Robotics, Hardware Design, Mechatronics Systems
Abstract: Mammalian entertainment robots are popular for their friendly appearance, yet few achieve true legged locomotion due to high motor torque and power consumption requirements. In this study, we propose a mammalian electric quadruped walking robot employing Gravitationally Decoupled Actuation (GDA), which enables high-speed locomotion while minimizing the power required for walking. Using a dynamics simulator, we compared two configurations—conventional rotary joint actuation and the GDA model—and evaluated their motor torque and power requirements. Furthermore, based on commercial actuator specifications, we examined the feasibility of a practical GDA-based design. We also investigated a gravity compensation mechanism aimed at reducing motor peak torque and enabling long-duration operation. Simulation results showed that the GDA configuration reduced torque and power demand, achieving approximately 2.24 times the walking speed of the rotary joint actuation model under the same power limit. Introducing gravity compensation further reduced peak torque during the stance phase, making the system more suitable for prolonged operation.

Keywords: Robotics, Decision-making systems
Abstract: This paper addresses an energy-efficient object collection problem using two mobile robots twins, equipped with a flexible tool such as a net or tether. This problem is subject to a routing problem for a collection sequence that minimizes total energy consumption of both robots for collecting scattered, multiple objects. To this end, we propose a dynamic programming-based route planning method, where the robot state is defined by energy consumption and its position. As the flexible tool allows the robots to collect a number of objects all at once, we implement a grouping circle of the tool, the radius of which depends on the number of the collected objects. The method incorporates a load-dependent velocity model as well as the load-dependent radius of object grouping circles to realistically reflect robot dynamics during the cooperative collection and transportation of the objects. Simulations were conducted under various scenarios with different object distributions, start positions of the robots, and the maximum velocities of the robots. The simulation results demonstrate that the proposed method reduces both energy consumption and travel time compared to baseline methods.

Keywords: Renewable and sustainable energy, Hardware Design, Integration Platforms
Abstract: In an ever-adapting world, renewable energy resources are becoming increasingly relied upon for modern life. The rapidly expanding industry of wind energy, which has come to dominate this sector in production and versatility, suffers from turbines which face catastrophic blade failures and high maintenance costs. Many new projects are being developed to prevent, detect, and mitigate damage to turbine components to combat these issues. While there are several tools which actively determine turbine blade damage, there are currently no widely used systems to monitor the blades remotely and consistently during operation. This project proposes an inexpensive strain measurement array, comparable to lab equipment, which can collect data for the inverse finite element method and detect damage. The basic system designed could be incorporated into a turbine’s supervisory control and data acquisition to decrease turbine downtime, improve operational safety, and extend the lifespan of wind turbine blades, reducing maintenance costs and improving sustainability within the wind energy industry.

Keywords: Software Design, Robotics, Decision-making systems
Abstract: In this study, we propose a grid-based marking prompt framework to enhance spatial understanding in vision-language models (VLMs). The framework integrates object detection, background masking, and number overlaying, to enable VLMs to interpret spatial and contextual instructions more effectively. By inputting numbered images along with natural language instructions, the VLM selects the number corresponding to the most semantically appropriate location. The framework operates without requiring prior information such as 3D models or physical markers. Moreover, the proposed framework allows flexible rule adaptation through prompt engineering alone, providing general applicability across various objects and environments.

We conducted two experiments for the object placement task. In experiment 1, shelf images captured by a service robot were used to evaluate the placement selection accuracy of VLM. In experiment 2, the framework was implemented on a service robot and conducted the object placement task at positions selected by the VLM in a real-world environment. The framework achieved a high success rate in both experiments, demonstrating the effectiveness and practical utility of the framework in real-world environments.

Keywords: Software Design, Integration Platforms, Robotics
Abstract: In this paper, we describe the automated simulation and reporting pipeline implemented for our autonomous racing stack, ur.autopilot. The backbone of the simulation is based on a high-fidelity model of the vehicle interfaced as a Functional Mockup Unit (FMU). The pipeline can execute the software stack and the simulation up to three times faster than real-time, locally or on GitHub for Continuous Integration/Continuous Delivery (CI/CD). As the most important input of the pipeline, there is a set of running scenarios. Each scenario allows the initialization of the ego vehicle in different initial conditions (position and speed), as well as the initialization of any other configuration of the stack. This functionality is essential to validate efficiently critical modules, like the one responsible for high‑speed overtaking maneuvers or localization, which are among the most challenging aspects of autonomous racing. Moreover, we describe how we implemented a fault injection module, capable of introducing sensor delays and perturbations as well as modifying outputs of any node of the stack. Finally, we describe the design of our automated reporting process, aimed at maximizing the effectiveness of the simulation analysis.

Keywords: Robotics, Hardware Design
Abstract: This paper presents a simulation-based evaluation of BIWAKO-8, a compact autonomous surface robot equipped with a transformable locomotion mechanism that switches between the Station-keeping mode and the Straight-travel mode. A Webots-based simulation environment replicating field conditions was used to compare the two modes under four settings: ideal, disturbance-only, localization-error-only, and a realistic combination. Across all settings, the Straight-travel mode completed missions faster, whereas the Station-keeping mode consistently consumed less energy, revealing a clear speed–energy trade-off. Disturbances had negligible impact on the energy efficiency of the Straight-travel mode, while localization error increased energy use in both modes by inducing more frequent heading corrections. These findings support hybrid operation—using the Straight-travel mode for transit and the Station-keeping mode for precise positioning—as an effective strategy for long-duration autonomous surface missions.

Keywords: Integration Platforms, Robotics, Large-scale Infrastructure Systems
Abstract: Stereo matching between different cameras presents significant challenges, particularly for wide-area scanning systems using pan-tilt cameras with continuously changing orientations. This paper proposes a 3-D model-based multi pan-tilt camera system that establishes stereo correspondence between cameras by introducing a shared 3-D coordinate space derived from a reference model. Our approach reduces matching complexity by leveraging 3-D model-based calibration, enabling scalable volumetric capture for real-world measurement and monitoring applications. To validate the proposed system, we conducted wide-area experiments on outdoor Heating, Ventilation, and Air Conditioning (HVAC) units. Each pan-tilt camera captured 25 high-resolution images, stitched into panoramic mosaics. Using 3-D reference points from a laser scanner, we established projective mappings through feature-point correspondence registration and mapped pixel coordinates to pan-tilt angles via linear interpolation. The experimental results demonstrated that multiple camera views can be directly aligned through their shared 3-D model, achieving stereo correspondence without appearance-based matching. Volumetric measurement through stereo DIC on 4000 frames captured at 400 fps successfully show 3-D vibration with sub-millimeter velocity accuracy and identified dominant frequencies, validating the system's effectiveness for large-scale structural monitoring and industrial diagnostics.

Keywords: Decision-making systems, Robotics
Abstract: Wheeled robots are increasingly utilized for environmental exploration and data collection in environments inaccessible to humans, such as the lunar surface. While adaptive sampling methods like the Upper Confidence Bound (UCB) can balance the trade-off between exploration and measurement data exploitation to model Regions of Interest (ROI), they do not explicitly incorporate robot traversability into the trade-off. Consequently, when faced with multiple scientifically promising locations, the UCB may select paths for the robot that are costly in terms of travel time, leading to time-inefficient surveys, particularly in sloped terrain. To address this limitation, we propose the Slope-aware Upper Confidence Bound (SaUCB), a novel acquisition function that integrates a traversability score based on terrain slope directly into the decision-making process. This allows the robot to explicitly balance exploration, exploitation, and traversal time. Through an extensive simulation study in realistic geological features of lunar terrain models, we demonstrate that the proposed approach generates significantly more time-efficient survey paths. Our method also demonstrates an improved trade-off between investigation time and modeling accuracy within the ROI compared to conventional approaches.

Keywords: Robotics, Hardware Design
Abstract: We developed Roller-Walker II, a leg-wheeled hybrid robot that primarily uses 3D-printed plastic as its structural material. This robot achieves both walking and wheel-based locomotion called “Roller-Walk” by rotating its ankle joints 90 degrees to switch between different foot configurations. The ankle switching mechanism also employs 3Dprinted lead screws, realizing a compact, lightweight, and high load-bearing capacity mechanical configuration. Locomotion experiments were conducted for both walking and Roller-Walk modes, confirming that the robot can successfully propel itself in both modes without issues. Particularly in Roller-Walk mode, the robot achieved a maximum velocity of 3.43 m/s and a CoT of 0.14, placing it among the group of walking robots with the highest mobility performance when compared to other walking robots.

Keywords: Hardware Design, Robotics, Mechatronics Systems
Abstract: This research quantitatively evaluates the energy-saving effects of lightweighting the structural parts of industiral robots by substituting conventional aluminum alloys with plastics. We prototyped three types of 3-DOF vertically articulated robots with structural parts made of A5052 aluminum alloy, a carbon fiber non-woven composite (Feltcarbon), and a 3D-printed composite (Onyx+CF). We then measured the energy consumption of each joint and the total manipulator during a 10-cycle pick-and-place task. The results show that replacing the aluminum alloy with Feltcarbon and Onyx+CF reduces the manipulator's total energy consumption by 25% and 30%, respectively. At the single-joint level, maximum energy savings reached 44% for Feltcarbon and 53% for Onyx+CF compared to the A5052 baseline.

Keywords: Hardware Design, Robotics
Abstract: 3D printing technology offers a promising approach for fabricating lightweight and low cost versions of traditionally heavy and expensive components, such as gear reducers. While previous work exists on plastic based reducers, challenges related to low stiffness remain. In this paper, we designed and prototyped a trochoidal gear reducer primarily constructed from a high strength potassium titanate fiber-reinforced material. We conducted a comparative evaluation of two material variants, the nylon based NTL34M and the more stiff PPS based RT4. The experimental results showed no significant difference between the two materials such as no-load running current, joint stiffness, and static torque efficiency. However, in dynamic torque efficiency measurement, the RT4 reducer outperformed the NTL34M version.

Keywords: Hardware Design, Robotics, Renewable and sustainable energy
Abstract: 3D printed parts are widely used, which often require high strength and stiffness. However, 3D printed parts face various challenges, properties for printing materials, continuous fiber orientation, moisture absorption, and printing accuracy. Quantitative evaluation of each of these factors will determine appropriate design methods. Therefore, some studies are currently underway to clarify their properties. One example is the insertion of continuous fibers into 3D printed parts, which has been shown to improve strength and stiffness. However, there has been little quantitative evaluation of the orientation　and amount of continuous fibers insertion. Especially, while the insertion position is theoretically critical to strength and stiffness, there has been no prior approach evaluating the effect of CF layer insertion position. In this study, we quantitatively evaluated the degree of strength and stiffness improvement by inserting continuous carbon fiber into 3D printed parts during printing. To evaluate the insertion orientation, amount and location, we calculated theoretical equations and discussed the discrepancy between theoretical value and experimental results. The results showed that the degree of stiffness improvement achievable in 3D printed parts was clarified in comparison with theoretical values.

Keywords: Hardware Design, Robotics
Abstract: In recent years, the demand for industrial robots has been increasing year by year and is expected to continue to grow in the future．On the other hand, efforts toward the Sustainable Development Goals (SDGs) and carbon neutrality are spreading worldwide．This paper aims to develop a lightweight reduction gear for energy-saving robots. To reduce the weight of robots, it is necessary to replace conventional metal materials with new lightweight materials. However, weight reduction generally reduces rigidity, which tends to lower the high-speed and high-precision performance required of industrial robots. The authors have studied the possibility of replacing metal parts with machined CFRP, POM, and 3D printer resin parts, and have examined the adaptability of these parts. In the previous report, we experimentally compared the weight, no-load running torque, torque–torsional characteristic, and dynamic torque transfer efficiency of 20 combinations of metal and resin reduction gears. In this report, we focus on the yield strength of materials and analyze the load distribution, stress of reduction gear components, taking into account the effects of assembly and fabrication errors. We then calculate and compare the output torque per volume and weight of four types of reduction gears using metal and resin combinations. Under the condition of designing within the elastic range of the materials, it was demonstrated that using resin gears could increase the maximum output while saving space and reducing weight. Additionally, the validity of the theoretical values was verified by comparing them with FEM analysis results.

Keywords: Hardware Design, Robotics
Abstract: The authors have been researching and developing snake robots that can move over fragile ground. Real desert snakes can move over sand using sidewinding locomotion. To realize this locomotion by a snake robot with three dimensions movements, many active DOFs are required to lift its body segments. In this study, we focus on the regularity of sidewinding locomotion passively and report the design of a new mechanism to lift the body segments in conjunction with active degrees of freedom in the horizontal direction. We explain about analysis of the Snake Robot's Trajectory Planning and Sidewinding Movement.

Keywords: Integration Platforms, Software Design, Decision-making systems
Abstract: We propose a system that offloads hand-waving action recognition processing in service robots to a field-programmable gate array (FPGA) accelerator. The accelerator implements a LUTNet-based reservoir computing (LUTNet-RC) model. Conventionally, such processing was performed by laptops connected to robots. Nevertheless, the proposed system offloads some of these task processing to an FPGA. This improves the overall efficiency of the robot and optimizes the use of computational resources. The experimental results show that in hand-waving action recognition, the LUTNet-RC on FPGA achieves higher accuracy and precision than the conventional echo state network (ESN) model, while significantly reducing the power consumption. In addition, the system can be expanded to large-scale networks, which are challenging to implement on laptops.

Keywords: Human-robot Interaction / Collaboration, Robotics, Assistive Robotics
Abstract: This study presents a library service robot rapidly developed using cloud artificial intelligence (AI) services and empirically demonstrates the robot’s service in a public library. The developed robot system interacts with a cloud service for speech recognition and synthesis with high accuracy and low latency over a short duration. The robot has been in operation for three years in a university library. To conduct demonstration experiments at a public library, a book search system that does not use personal information was used, which differs from the system that has been in operation at the university library for several years. In addition, for the demonstration experiment at the public library, we formally negotiated with the library and set conditions such as the questioning method for the verification experiment. We show the feasibility of the proposed system with a spoken dialogue service robot by conducting demonstration experiments in the university and public libraries.

Keywords: Assistive Robotics, Human-robot Interaction / Collaboration, Decision-making systems
Abstract: Falls are a leading cause of injury in older adults, requiring detection systems that are both sensitive and reliable. We present a multi-modal robotic framework that integrates wearable sensing, vision-based verification, and dialogue-driven assessment. A smartwatch streams inertial data, with thresholds tuned through pilot testing to maximise fall sensitivity. Vision verification is performed using a fine-tuned YOLOv11 model, while Whisper ASR and a lightweight GPT-based classifier enable simple verbal checks of user responsiveness. Our tuned thresholds outperformed published baselines (F1 = 0.857), and the vision module achieved strong accuracy (mAP@0.5 = 0.827). In integrated trials, the system reached a 90.6% success rate with a mean end-to-end response time of 43.5 seconds. These results show that combining complementary modalities enhances robustness and moves socially assistive robots toward interactive fall response in real-world care.

Keywords: Robotics, Machine Learning
Abstract: Mobile robot navigation in dynamic environments that contain pedestrians is one of the key challenges in the development of autonomous mobile service robots. This field, known as social navigation, has seen significant research progress using reinforcement learning approaches. In recent years, numerous diffusion-based reinforcement learning methods capable of generating diverse actions have been proposed. However, compared to conventional reinforcement learning approaches, the diffusion model's slow generation process presents a significant barrier to real-time processing. To address this, we propose a method for knowledge distillation of conditional diffusion models by combining Gaussian Prior with Flow Matching to enable faster action generation in dynamic environments. Experiments using a crowd navigation benchmark in simulation environments demonstrate that a significant reduction of the time required for action generation is possible while maintaining nearly the same performance as teacher models.

Keywords: Human Factors, Machine Learning, Assistive Robotics
Abstract: Work-related musculoskeletal disorders (WMSDs) persist in material-handling jobs where lifting, twisting, and carrying induce high, localized muscle demands. This paper presents a sensor-driven framework that (i) detects biomechanical strain from surface electromyography (sEMG) and (ii) quantifies the impact of a passive back-support exoskeleton during industrially relevant tasks. With data from 20 participants performing standardized tasks with and without the device, we introduce a data-driven strain labeling method that replaces ad-hoc thresholds with piecewise linear regression to identify individualized strain onset. A compact deep neural network handles severe class imbalance via SMOTE and decision-threshold optimization, yielding 83.5% overall accuracy and a macro-averaged F1-score of 0.70 for binary strain classification. Muscle-specific analyses reveal significant reductions in biceps and oblique activation (p < 0.001) alongside compensatory increases in erector spinae and lower-limb activity, indicating load redistribution rather than uniform offloading. The result is a scalable, real-time approach that captures both when strain begins and how effort shifts across muscle groups, capabilities that traditional peak-sEMG or subjective assessments miss. By uniting wearable sensing, automated strain onset detection, and imbalance-aware learning, this work advances objective, continuous, and human-centered ergonomic monitoring and provides actionable evidence for the deployment of passive exoskeletons in smart industrial environments.

Keywords: Assistive Robotics, Rehabilitation Systems
Abstract: Accurate estimation of ankle joint impedance during walking is crucial for understanding human gait dynamics. However, conventional methods often rely on large experimental setups or invasive techniques, which limit practicality and accessibility. This study aimed to develop and validate a compact method for applying perturbations to the ankle joint using Treadmill Belt Deceleration. An external torque was applied to the ankle joint by rapidly decelerating the treadmill belt without prior notice to the participant. A differential model was employed to estimate joint impedance for two participants, with stiffness and viscosity normalized by body weight. Only models with stable poles (pole magnitude < 1) were considered. The resulting impedance estimates aligned with expected gait-phase trends, confirming the method’s consistency and physiological relevance. This approach enable experiments to be conducted in more compact settings than those in previous studies involving large walkways. These results are comparable with previous findings, demonstrating the validity of the proposed method. This method offers a practical solution for conducting gait and balance studies in constrained spaces.

Keywords: Human-robot Interaction / Collaboration, Machine Learning, Robotics
Abstract: We present a fully online framework for streaming human upper-limb kinematics estimation from a single 360-degree camera. Incoming frames are processed sequentially through vertical-boundary-aware tracking, pseudo-perspective rendering, and Neural Localizer Fields to estimate a sparse set of 3D anatomical landmarks in real time. These landmarks are mapped to an OpenSim-compatible biomechanical model, with joint angles computed on the fly via an online inverse kinematics solver. The system achieves end-to-end latencies as low as 22.9 ms on a high-performance setup. Evaluated in a single-participant scenario involving an initial T-pose calibration and repeated object displacement toward the camera, it demonstrates robust performance under moderate self-occlusion and spherical distortion. While tested in a constrained setting, its modular, real-time design makes it a promising candidate for human–robot interaction and other motion analysis applications, enabling minimal, markerless, and anatomically interpretable upper-limb tracking from omnidirectional vision.

Keywords: Robotics, Human-robot Interaction / Collaboration, Assistive Robotics
Abstract: In this research, we focus on the design and applications of a 12DoF(degrees of freedom) dual-arm inflatable robot. The design utilizes a combination of soft bodies, such as inflatable links, and hard joints as servo actuators. This combination enables us to retain the passive safety and compliance of soft robots while leveraging the control accuracy of hard joints. First, we built a 6DoF inflatable arm with a gripper (1DoF). Then, we replicated the single arm to construct a 12DoF dual-arm robot. We then examine the feasibility of this robot for several contact-based tasks, like hammering a nail, pick and place of a handkerchief, and dressing assistance. For our evaluation, we report successful trials and servo faults observed during experiments, as verified by video recordings. With the tasks shown, we confirm the robot's inherent compliance and showcase the qualitative report for the feasibility of contact-rich tasks based on a small number of trials.

Keywords: Machine Learning, Hardware Design, Decision-making systems
Abstract: Simulation plays a crucial role in optimizing complex system designs. While individual subsystems may have relatively short processing times, when the entire system is integrated, the significant time and computational resources required remain a considerable challenge. On the other hand, system integration introduces effects that cannot be easily formalized, making integrated simulation indispensable. To address this challenge, this study investigated the effectiveness of improving accuracy by adding training data in sequential approximate optimization (SAO) using Integration Neural Networks (INN). INN maintains high accuracy even with limited training data by integrating neural networks that handle phenomena that can be formulated and neural networks that handle phenomena that cannot be formulated. In conventional SAO, the initial training data set is determined based on a rule called “Uncle Barney's rule. In contrast, this study hypothesized that optimization with fewer simulation runs could be achieved by strategically replacing random, low-quality initial training data with data selected based on intermediate learning results. Specifically, we investigated whether the total amount of training data required could be reduced by decreasing the number of initial training data points and increasing the number of data points added during the optimization process. The results of the experiments showed that stable solutions within the error range were obtained for all cases. Furthermore, it was demonstrated that a balance between data reduction and optimization stability could be successfully achieved by setting the number of initial training samples to three times the number of design variables (21 samples). As a result, the number of simulation runs required to construct the training dataset was reduced by approximately 50.2% compared to the conventional SAO method, demonstrating significant potential for shortening the design cycle in product optimization.

Keywords: Machine Learning, Robotics, Integration Platforms
Abstract: The rapid expansion of resource-constrained mobile platforms, such as mobile robots, wearable systems, and Internet-of-Things devices, has heightened the need for computationally efficient neural network controllers (NNCs) that can operate within strict hardware limitations. Although deep neural networks (DNNs) achieve high performance in control applications, their considerable computational complexity and memory demands hinder practical deployment on edge devices. This study presents a comprehensive model compression methodology that employs component-aware structured pruning to determine the optimal pruning magnitude for each group, thereby balancing compression and stability for NNC deployment. The proposed approach is rigorously evaluated on Temporal Difference Model Predictive Control (TD-MPC), a leading model-based reinforcement learning algorithm, with systematic integration of mathematical stability guarantees, specifically Lyapunov criteria. The principal contribution is a principled framework for identifying the theoretical limits of model compression while preserving controller stability. Experimental results indicate that the methodology effectively reduces model complexity while maintaining essential control performance and stability. Additionally, the approach defines a quantitative boundary for safe compression ratios, enabling practitioners to systematically determine the maximum permissible model reduction before compromising critical stability properties, thus supporting the reliable deployment of compressed NNCs in resource-limited environments.

Keywords: Robotics, Machine Learning, Decision-making systems
Abstract: Attitude and Heading Reference Systems (AHRSs) based on Error-State Extended Kalman Filters (ES-EKF) require careful tuning of stochastic parameters to achieve optimal performance. Traditional approaches rely on Allan Variance (AV) analysis for parameter identification, which, while physically grounded, does not guarantee optimal navigation performance. This paper presents three advanced optimization methodologies—Gaussian Processes (GP), Nondominated Sorting Genetic Algorithm III (NSGA-III), and Multi-Objective Tree-structured Parzen Estimator (MO-TPE)—to systematically fine-tune critical ES-EKF parameters directly against flight performance metrics. The optimization targets accelerometer/magnetometer noise/bias characteristics, vehicle dynamics/Earth magnetic field uncertainty, and Innovation Filter (IF) thresholds across a multi-dimensional parameter space. Experimental validation using 19 real Unmanned Aerial Vehicle (UAV) flights demonstrates substantial improvements over traditional AV-based tuning. The proposed methods achieve 81-91% reductions in Root Mean Square (RMS) attitude errors, 82-91% improvements in Mean Absolute Errors (MAE), and 41-93% enhancements in estimation consistency across roll, pitch, and yaw axes. MO-TPE consistently delivers the best overall performance, achieving optimal results in 7 out of 9 metric-axis combinations, followed closely by NSGA-III, while GP provides competitive single-objective optimization. The largest improvements are observed in yaw estimation, where traditional approaches struggle most with magnetic disturbances. These results demonstrate that machine learning-driven parameter optimization can significantly enhance AHRS accuracy and robustness without modifying the underlying ES-EKF structure, offering a practical path for improving navigation performance in real-world UAV applications.

Keywords: Renewable and sustainable energy, Environment / Ecological Systems, Hardware Design
Abstract: With the aim of mitigating the impacts caused by variations in the rainy season that directly affect the production of forage grasses in San Lucas, Michoacán, this work presents the design and evaluation of a direct-solar pumping system for drip irrigation over an area of 0.74 ha. This development considered the meteorological data from the National Renewable Energy Laboratory database, as well as the hydraulic characteristics of the plot and the water requirements of the crop, in order to design both the hydraulic infrastructure and the photovoltaic array. The proposal was evaluated using the specialized software, System Advisor Model, resulting in a system capable of supplying a required flow rate of 3.52 lps during the dry season, ensuring continuous drip irrigation operation, without the use of battery banks, from 10:00 AM. to 4:00 PM. The obtained results support the technical and economic feasibility of the system, therefore, its implementation is recommended as a cost-effective and sustainable alternative for local agricultural production.

Keywords: Control Technologies, Robotics, Intelligent Transportation Systems
Abstract: Robot-based services are increasingly being integrated into everyday life. In the diverse environments in which humans live, systems are expected to respond to dynamically changing surroundings by enabling distributed control and cooperation. However, in spaces where various robots coexist, it is practically challenging to ensure that all robots in the space can communicate all necessary information with one another. This poses challenges related to communication failures, the need for unified communication standards, and legal regulations for robots that do not comply with these standards. Furthermore, from the perspective of coexistence with robots in urban areas, there are many entities, such as humans, birds, and cats, with which standardized communication is not feasible. When introducing robots into environments such as cities, it is necessary to share pathways such as roads, sidewalks, and corridors, as well as public infrastructure such as buses and trains, with non-communicating entities. Therefore, in the social application of such robots, it is desirable to construct a system that can operate on the premise that it cannot obtain complete information about other robots or living organisms, and it is necessary to consider a system that can continue its mission even when information about other robots is ambiguous.　 In this study, we investigated a distributed cooperative strategy planning algorithm for a multi-agent system with multiple tasks of patrolling three designated points specified by coordinates in an unknown space. The algorithm assigns tasks to each agent and enables the system to continue its mission even when agents cannot accurately obtain information about one another. We also verified the effectiveness of this algorithm using actual robots and report the results.

Keywords: Control Technologies, Automation, Software Design
Abstract: The development of distributed embedded controllers can be significantly enhanced by adopting a model-driven development approach benefiting from the adoption of low-code strategies, where Petri net-based modeling can support rigorous system specification, verification, automatic code generation and direct deployment. This paper presents a development flow where, starting from the global system model expressed through an IOPT Petri net model, a decomposition strategy allow to obtain a set of concurrent sub-models. These sub-models are associated with different time domains in terms of execution time and deployed into a set of networked controllers. Several common protocols can be selected to support inter-controller communication ensuring synchronization between the networked controllers, namely I²C, UART and TCP/MQTT. The model-driven development approach is supported by the IOPT-Tools framework, a web-based platform freely available at http://gres.uninova.pt/IOPT-Tools/.

Keywords: Network Systems, Integration Platforms, Human-robot Interaction / Collaboration
Abstract: This study proposes an automatic pose estimation method for distributed RGB-D cameras aimed at supporting the configuration of intelligent spaces. The proposed method is based on the independent pose estimation of each camera, that is effective for scalability of intelligent space configuration. The distributed camera poses are automatically estimated by utilizing the extrinsic parameters of already aligned cameras and the point cloud alignments among the adjacent cameras and the 3D map. This approach significantly reduces the workload required for building intelligent spaces. Furthermore, a person tracking system is constructed as the application, demonstrating the practical usability of the proposed method.

Keywords: Intelligent Transportation Systems, Machine Learning, Robotics
Abstract: The extraction and matching of the feature in V-SLAM is important to ensure the accuracy of the location. This paper presents a dynamic feature point filtering algorithm which combines the semantic segmentation and the geometric constraints. The algorithm performs the semantic segmentation on the preprocessed RGB images after highlight/shadow removal to initially obtain the masks of the suspected dynamic objects. Then the motion consistency detection is integrated to determine the motion states of the feature points, preserving the static features while filtering the dynamic ones, while the dealt features are subsequently used for camera motion matrix estimation. Experimental have been performed to validated on the public dataset, i.e. TUM, KITTI and custom-built dataset to demonstrate the effectiveness in V-SLAM systems.

Keywords: Human-robot Interaction / Collaboration, Software Design
Abstract: The problem of human isolation and loneliness is growing worldwide. Although there are ways to provide human support, there are sustainability issues, so support using conversational physical partners, such as robots, is attracting attention as an alternative solution. Most previous attempts have been based on a single robot, and there are few attempts that use multiple robots, especially those that form a family and try to save the user from isolation and loneliness. This study aims to develop a basic infrastructure system to realize such a family consisting of multiple robots and a single human. More specifically, in this paper we report on the implementation and preliminary verification of two types of dialogue generation for robots using ROS (Robot Operating System) and GPT-4o-mini: one is a centralized system (in which the command center generates conversation scenarios for each robot in batches) and the other is a distributed system (in which each robot generates conversations individually).

Keywords: Entertainment and Educational Systems, Robotics, Software Design
Abstract: This work presents the design and development of BEATRIXsim, a simulation environment for integration into the educational BEATRIX robot platform. BEATRIXsim is implemented in CoppeliaSim due to its support for kinematics and low computational load. Both the physical robot (BEATRIX) and digital robot (BEATRIXsim) are interfaced using an Arduino Mega microcontroller and the Robot Operating System (ROS) using the rosserial protocol. Custom-built ROS messages, services and a bridge are developed for synchronised operation, communication, control and real-time feedback of both the physical and digital robots. This simulation environment includes a Graphical User Interface (GUI) to allow users to simultaneously control both the physical hardware and the digital robot, but also their individual control. BEATRIXsim serves as a tool for experimentation and testing processes without the need for its physical counterpart, reducing risks, costs and development time. Overall, this work shows that an affordable humanoid robot for education purposes can be enhanced by the use of digital technology, providing a robust platform for real-time control, experimentation and teaching in robotics.

Keywords: Robotics, Human-robot Interaction / Collaboration, Mechatronics Systems
Abstract: Legacy (unsupported) robotic platforms often lose research utility when manufacturer support ends, preventing integration of modern sensing, speech, and interaction capabilities. We present the textit{Enhanced NAO}, a revitalized version of Aldebaran's NAO robot featuring upgraded beamforming microphones, RGB-D and thermal cameras, and additional compute resources in a fully self-contained package. This system combines cloud-based and local models for perception and dialogue, while preserving the NAO’s expressive body and behaviors. In a pilot user study validating conversational performance, the Enhanced NAO delivered significantly higher conversational quality and elicited stronger user preference compared to the textit{NAO AI Edition}, without increasing response latency. The added visual and thermal sensing modalities established a foundation for future perception-driven interaction. Beyond this implementation, our framework provides a platform-agnostic strategy for extending the lifespan and research utility of legacy robots, ensuring they remain valuable tools for human-robot interaction.

Keywords: Robotics, Control Technologies, Automation
Abstract: In torque-controlled robotic systems, external disturbances, whether from physical interactions or unmodeled dynamics, can significantly affect stability and performance. This paper investigates two different approaches for external disturbance compensation within closed-loop torque control frameworks: passivity-based compensation and residual-based compensation. The former leverages discrepancies between predicted and measured joint torques to estimate external forces, enabling fast reactive control. The latter relies on passivity theory to guarantee stability by shaping the system’s energy exchange, making it robust to modeling uncertainties and sensor noise. We present the control formulations of both methods and evaluate them through experimental benchmarks in robotic tasks requiring compliant interaction. The theoretical and experimental results highlight the distinct strengths and limitations of each strategy in terms of responsiveness, stability, robustness, offering guidelines for selecting an appropriate disturbance compensation method depending on the application scenario.

Keywords: Robotics, Automation
Abstract: In recent years, the number of farm households in Japan has been decreasing due to the aging of farmers and the shift away of young labors, thus automation and labor saving are required as a solution. In this study, we propose an automatic harvesting robot for the purpose of automating harvesting, one of the tasks in Citrus unshiu cultivation. Previously, we proposed an end effector consisting of a suction pad and scissors for automatic harvesting, one of the tasks in Citrus unshiu cultivation. In this paper, we report a system integration of a crawler-type mobile manipulator that mount an arm robot with the end effector and automatically travel on slopes and uneven terrain in the field, also a method for recognizing Citrus unshiu using the Watershed algorithm that can separate and recognize the adjacent area of the target. This system is expected to enable night harvesting, which is expected to contribute to the sustainable development of agriculture.

Keywords: Machine Learning, Robotics, Automation
Abstract: Imitation learning has gained significant attention as a promising approach to enable flexible and generalizable robot motion generation across diverse tasks. However, existing models often suffer from long inference times, limiting their applicability to high-speed and fine-grained tasks. Moreover, while faster computation enables shorter inference cycles, it introduces new challenges such as overcontrol and motion instability when the inference frequency exceeds the sensor sampling rate. To address these issues, we propose QDM-RNN, a lightweight motion generation model that learns from slow but high-quality demonstrations and remains robust under high-frequency inference. Our method utilizes Softmax Transformation, which discretizes the robot end-effector pose into a high-resolution probability distribution, enabling accurate and smooth trajectory prediction. Furthermore, by incorporating multi-timestep prediction, which simultaneously predicts several future steps, our model mitigates instability arising from the mismatch between sensor and inference rates, ensuring consistent long-horizon motion generation. We validated the effectiveness of our approach through real-world robotic experiments on a Sport Stacking task, which requires both high speed and precision. QDM-RNN maintained high success rates and motion stability at speeds up to seven times faster than demonstrations, and showed robustness to external disturbances including background changes, lighting variations, dynamic object perturbations, and obstacles.

Keywords: Human Factors, Human-robot Interaction / Collaboration, Machine Learning
Abstract: For practical deployment of autonomous robots in human-existed environment, it is essential to train robot policies with simulation environments that reflect the diversity of human body features and motion behaviors. However, existing datasets often overlook this need, relying on simplified skeleton models that ignore variations in age, height, or body shape. Moreover, realistic scenarios representing such diversity are largely missing due to the high cost and complexity of data collection. In this work, we address these limitations by constructing a human motion dataset that captures a wide range of body types and age groups, using accurate and characterized body models. These detailed representations allow robots to better learn how physical attributes influence movement, thereby enhancing their responsiveness and safety during interaction. To further expand the dataset efficiently, we also explore data generation techniques that create diverse motion samples from limited inputs. Our approach enables the scalable construction of simulation environments that reflect human variability, offering a valuable resource for future robot policy learning.