Pub Date : 2024-11-18DOI: 10.1109/TCYB.2024.3490580
Shunyi Zhao;Tianyu Zhang;Yuriy S. Shmaliy;Xiaoli Luan;Fei Liu
Integrating the advantage of the unbiased finite impulse response (UFIR) filter into the Kalman filter (KF) is a practical yet challenging issue, where how to effectively borrow knowledge across domains is a core issue. Existing methods often fall short in addressing performance degradation arising from noise uncertainties. In this article, we delve into a Bayesian transfer filter (BTF) that seamlessly integrates the UFIR filter into the KF through a knowledge-constrained mechanism. Specifically, the pseudo marginal measurement likelihood of the UFIR filter is reused as a constraint to refine the Bayesian posterior distribution in the KF. To optimize this process, we exploit the Kullback-Leibler (KL) divergence to measure and reduce discrepancies between the proposal and target distributions. This approach overcomes the limitations of traditional weight-based fusion methods and eliminates the need for error covariance. Additionally, a necessary condition based on mean square error criteria is established to prevent negative transfer. Using a moving target tracking example and a quadruple water tank experiment, we demonstrate that the proposed BTF offers superior robustness against noise uncertainties compared to existing methods.
{"title":"Bayesian Transfer Filtering Using Pseudo Marginal Measurement Likelihood","authors":"Shunyi Zhao;Tianyu Zhang;Yuriy S. Shmaliy;Xiaoli Luan;Fei Liu","doi":"10.1109/TCYB.2024.3490580","DOIUrl":"10.1109/TCYB.2024.3490580","url":null,"abstract":"Integrating the advantage of the unbiased finite impulse response (UFIR) filter into the Kalman filter (KF) is a practical yet challenging issue, where how to effectively borrow knowledge across domains is a core issue. Existing methods often fall short in addressing performance degradation arising from noise uncertainties. In this article, we delve into a Bayesian transfer filter (BTF) that seamlessly integrates the UFIR filter into the KF through a knowledge-constrained mechanism. Specifically, the pseudo marginal measurement likelihood of the UFIR filter is reused as a constraint to refine the Bayesian posterior distribution in the KF. To optimize this process, we exploit the Kullback-Leibler (KL) divergence to measure and reduce discrepancies between the proposal and target distributions. This approach overcomes the limitations of traditional weight-based fusion methods and eliminates the need for error covariance. Additionally, a necessary condition based on mean square error criteria is established to prevent negative transfer. Using a moving target tracking example and a quadruple water tank experiment, we demonstrate that the proposed BTF offers superior robustness against noise uncertainties compared to existing methods.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 2","pages":"562-573"},"PeriodicalIF":9.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-18DOI: 10.1109/TCYB.2024.3486562
Tianbiao Shi;Fanglai Zhu
This article investigates the secure control problem for a class of Lipschitz nonlinear descriptor multiagent systems (MASs) with unknown inputs under Denial-of-Service (DoS) attacks. In order to address the presence of unknown state variables and external disturbances in both the state and output equations, a local unknown input observer (UIO) is developed for each follower agent. The proposed UIO is capable of simultaneously estimating the system state, measurement noise and unknown inputs through an interval observer. With regards to DoS attacks, we consider two types: those that maintain connectivity and those that paralyze it by disrupting the structure of the information communication topology graph. By utilizing the proposed UIO, a distributed compensation controller is designed to achieve asymptotic consensus for leader-following MASs under DoS attacks. Additionally, a comprehensive stability analysis of the closed-loop system is provided, taking into account switching systems. Finally, two simulation examples are presented to validate the effectiveness of the proposed UIO-based distributed secure control scheme.
{"title":"Distributed Secure Control for Nonlinear Descriptor Multiagent Systems With Unknown Inputs Under Denial-of-Service Attacks","authors":"Tianbiao Shi;Fanglai Zhu","doi":"10.1109/TCYB.2024.3486562","DOIUrl":"10.1109/TCYB.2024.3486562","url":null,"abstract":"This article investigates the secure control problem for a class of Lipschitz nonlinear descriptor multiagent systems (MASs) with unknown inputs under Denial-of-Service (DoS) attacks. In order to address the presence of unknown state variables and external disturbances in both the state and output equations, a local unknown input observer (UIO) is developed for each follower agent. The proposed UIO is capable of simultaneously estimating the system state, measurement noise and unknown inputs through an interval observer. With regards to DoS attacks, we consider two types: those that maintain connectivity and those that paralyze it by disrupting the structure of the information communication topology graph. By utilizing the proposed UIO, a distributed compensation controller is designed to achieve asymptotic consensus for leader-following MASs under DoS attacks. Additionally, a comprehensive stability analysis of the closed-loop system is provided, taking into account switching systems. Finally, two simulation examples are presented to validate the effectiveness of the proposed UIO-based distributed secure control scheme.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"472-485"},"PeriodicalIF":9.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670657","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-18DOI: 10.1109/TCYB.2024.3487934
Witold Pedrycz
Undoubtedly, machine learning (ML) has demonstrated a wealth of far-reaching successes present both at the level of fundamental developments, design methodologies and numerous application areas, quite often encountered in domains requiring a high level of autonomous behavior. Over the passage of time, there are growing challenges of privacy and security, interpretability, explainability, confidence (credibility), and computational sustainability, among others. In this study, we advocate that these quests could be addressed by casting them both conceptually and algorithmically in the unified environment augmented by the principles of granular computing. It is demonstrated that the level of abstraction, delivered by granular computing plays a pivotal role in the interpretation by quantifying the level of credibility of ML constructs. The study also highlights the principles of granular computing and elaborates on its landscape. The original idea of a comprehensive and unified framework of data-knowledge environment of ML is introduced along with a detailed discussion on how data and knowledge are used in a seamless fashion by invoking granular embedding and producing relevant loss functions. Key categories of knowledge-data integration realized at the levels of data and model (involving symbolic/qualitative models and physics-oriented models) and investigated.
{"title":"Granular Computing for Machine Learning: Pursuing New Development Horizons","authors":"Witold Pedrycz","doi":"10.1109/TCYB.2024.3487934","DOIUrl":"10.1109/TCYB.2024.3487934","url":null,"abstract":"Undoubtedly, machine learning (ML) has demonstrated a wealth of far-reaching successes present both at the level of fundamental developments, design methodologies and numerous application areas, quite often encountered in domains requiring a high level of autonomous behavior. Over the passage of time, there are growing challenges of privacy and security, interpretability, explainability, confidence (credibility), and computational sustainability, among others. In this study, we advocate that these quests could be addressed by casting them both conceptually and algorithmically in the unified environment augmented by the principles of granular computing. It is demonstrated that the level of abstraction, delivered by granular computing plays a pivotal role in the interpretation by quantifying the level of credibility of ML constructs. The study also highlights the principles of granular computing and elaborates on its landscape. The original idea of a comprehensive and unified framework of data-knowledge environment of ML is introduced along with a detailed discussion on how data and knowledge are used in a seamless fashion by invoking granular embedding and producing relevant loss functions. Key categories of knowledge-data integration realized at the levels of data and model (involving symbolic/qualitative models and physics-oriented models) and investigated.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"460-471"},"PeriodicalIF":9.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-18DOI: 10.1109/TCYB.2024.3491582
Hui Yu;Liqian Dou;Xiuyun Zhang;Jinna Li;Qun Zong
This article addresses the collision avoidance and formation control problem for multisatellite systems. A novel safe reinforcement learning (RL) algorithm based on an adaptive dynamic programming framework is proposed. The highlights of the algorithm are the adaptive distance-varying learning method to integrate online data with historical data and the usage of the barrier function (BF) to achieve collision avoidance. First, the BF is introduced into the designed cost function such that the multisatellite formation system can achieve obstacle avoidance and guarantee the safety. Next, a safe RL algorithm is developed through the critic network structure. A distance-varying weight is introduced, which combines experience replay samples with extrapolation samples. By minimizing the cost function, the optimal formation control policy can be obtained with an adaptive formation and self-learning ability. Then, the stability and safety of the proposed algorithm are analyzed. Finally, the effectiveness and superiority of the proposed algorithm are verified by numerical simulations.
{"title":"Safe Reinforcement Learning: Optimal Formation Control With Collision Avoidance of Multiple Satellite Systems","authors":"Hui Yu;Liqian Dou;Xiuyun Zhang;Jinna Li;Qun Zong","doi":"10.1109/TCYB.2024.3491582","DOIUrl":"10.1109/TCYB.2024.3491582","url":null,"abstract":"This article addresses the collision avoidance and formation control problem for multisatellite systems. A novel safe reinforcement learning (RL) algorithm based on an adaptive dynamic programming framework is proposed. The highlights of the algorithm are the adaptive distance-varying learning method to integrate online data with historical data and the usage of the barrier function (BF) to achieve collision avoidance. First, the BF is introduced into the designed cost function such that the multisatellite formation system can achieve obstacle avoidance and guarantee the safety. Next, a safe RL algorithm is developed through the critic network structure. A distance-varying weight is introduced, which combines experience replay samples with extrapolation samples. By minimizing the cost function, the optimal formation control policy can be obtained with an adaptive formation and self-learning ability. Then, the stability and safety of the proposed algorithm are analyzed. Finally, the effectiveness and superiority of the proposed algorithm are verified by numerical simulations.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"447-459"},"PeriodicalIF":9.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article investigates the observer-based human-in-the-loop (HiTL) optimal output cluster synchronization control problem for nonlinear multiagent systems (MASs). First, the leader is designed to be nonautonomous, with the unknown time-varying input monitored by the human operator directly. To address the problem that leader’s output is not available to each follower, an observer is designed. This observer features practical prescribed-time convergence, and independence of prior knowledge of leader’s input. Then, an augmented system consisting of observer dynamics and follower dynamics is constructed and a cost function is formulated. Accordingly, the HiTL optimal output cluster synchronization control problem is transformed into a solution to the Hamilton-Jacobian–Bellman equation (HJBE). Subsequently, the off-policy reinforcement learning algorithm is utilized to learn the solution to HJBE without complete knowledge of the system dynamics. To alleviate computational burden, the single critic neural network (NN) is employed for the algorithm implementation, with the least square method applied for training the NN weights. Finally, the simulation results are presented to verify the validity of the designed control scheme.
{"title":"Observer-Based Human-in-the-Loop Optimal Output Cluster Synchronization Control for Multiagent Systems: A Model-Free Reinforcement Learning Method","authors":"Zongsheng Huang;Tieshan Li;Yue Long;Hongjing Liang","doi":"10.1109/TCYB.2024.3490602","DOIUrl":"10.1109/TCYB.2024.3490602","url":null,"abstract":"This article investigates the observer-based human-in-the-loop (HiTL) optimal output cluster synchronization control problem for nonlinear multiagent systems (MASs). First, the leader is designed to be nonautonomous, with the unknown time-varying input monitored by the human operator directly. To address the problem that leader’s output is not available to each follower, an observer is designed. This observer features practical prescribed-time convergence, and independence of prior knowledge of leader’s input. Then, an augmented system consisting of observer dynamics and follower dynamics is constructed and a cost function is formulated. Accordingly, the HiTL optimal output cluster synchronization control problem is transformed into a solution to the Hamilton-Jacobian–Bellman equation (HJBE). Subsequently, the off-policy reinforcement learning algorithm is utilized to learn the solution to HJBE without complete knowledge of the system dynamics. To alleviate computational burden, the single critic neural network (NN) is employed for the algorithm implementation, with the least square method applied for training the NN weights. Finally, the simulation results are presented to verify the validity of the designed control scheme.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 2","pages":"649-660"},"PeriodicalIF":9.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article presents a novel online obstacle avoidance trajectory planning method for autonomous ground vehicles (AGVs) based on long short-term memory-attention (LSTM-Attention) networks. The proposed method can guide AGVs to perform emergency maneuvers when encountering sudden and moving obstacles, while also ensuring high levels of real-time performance and optimality. It consists of two parts: 1) offline training and 2) online planning. In the offline training phase, an AGV obstacle avoidance trajectory dataset is generated using numerical trajectory optimization methods to train the LSTM-Attention network. This training allows the network to capture the mapping between the relative information of the vehicle and the obstacles and the optimal control actions. The trained network is then used for online trajectory planning to achieve optimal feedback obstacle avoidance control for AGVs facing sudden obstacles. Furthermore, to address situations involving sudden obstacles in different directions and moving obstacles, a rotation coordinate system method is proposed, significantly expanding the application scenarios of the proposed approach. The effectiveness and real-time performance of the designed method are comprehensively validated through extensive simulation and physical experiments.
{"title":"Online Trajectory Planning Method for Autonomous Ground Vehicles Confronting Sudden and Moving Obstacles Based on LSTM-Attention Network","authors":"Zhida Xing;Runqi Chai;Kaiyuan Chen;Yuanqing Xia;Senchun Chai","doi":"10.1109/TCYB.2024.3486004","DOIUrl":"10.1109/TCYB.2024.3486004","url":null,"abstract":"This article presents a novel online obstacle avoidance trajectory planning method for autonomous ground vehicles (AGVs) based on long short-term memory-attention (LSTM-Attention) networks. The proposed method can guide AGVs to perform emergency maneuvers when encountering sudden and moving obstacles, while also ensuring high levels of real-time performance and optimality. It consists of two parts: 1) offline training and 2) online planning. In the offline training phase, an AGV obstacle avoidance trajectory dataset is generated using numerical trajectory optimization methods to train the LSTM-Attention network. This training allows the network to capture the mapping between the relative information of the vehicle and the obstacles and the optimal control actions. The trained network is then used for online trajectory planning to achieve optimal feedback obstacle avoidance control for AGVs facing sudden obstacles. Furthermore, to address situations involving sudden obstacles in different directions and moving obstacles, a rotation coordinate system method is proposed, significantly expanding the application scenarios of the proposed approach. The effectiveness and real-time performance of the designed method are comprehensively validated through extensive simulation and physical experiments.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"421-435"},"PeriodicalIF":9.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142637321","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article investigates a comprehensive data-driven event-triggered secure lateral control of autonomous vehicles under actuator attacks. We consider stabilization issues of autonomous vehicles subject to modeling difficulties, limited communication resources, and actuator attacks. The dynamic model decomposition (DMD) from data is exploited to characterize the inherent lateral dynamics model of autonomous vehicles, the event-triggered transmission scheme is utilized to alleviate communication burden for limited bandwidth network, and the sliding mode control scheme is designed to ensure the security of autonomous vehicles under actuator attacks. The stability analysis and the stabilization method as well as its algorithm are presented. The proposed secure control scheme can actively counteract the malicious effects caused by actuator attacks and integrates the advantages of both data-driven modeling and model-based control design. Finally, several comparative case studies show the effectiveness of the proposed secure control scheme.
{"title":"Data-Driven Event-Triggered Sliding Mode Secure Control for Autonomous Vehicles Under Actuator Attacks","authors":"Hong-Tao Sun;Xinran Chen;Zhengqiang Zhang;Xiaohua Ge;Chen Peng","doi":"10.1109/TCYB.2024.3490656","DOIUrl":"10.1109/TCYB.2024.3490656","url":null,"abstract":"This article investigates a comprehensive data-driven event-triggered secure lateral control of autonomous vehicles under actuator attacks. We consider stabilization issues of autonomous vehicles subject to modeling difficulties, limited communication resources, and actuator attacks. The dynamic model decomposition (DMD) from data is exploited to characterize the inherent lateral dynamics model of autonomous vehicles, the event-triggered transmission scheme is utilized to alleviate communication burden for limited bandwidth network, and the sliding mode control scheme is designed to ensure the security of autonomous vehicles under actuator attacks. The stability analysis and the stabilization method as well as its algorithm are presented. The proposed secure control scheme can actively counteract the malicious effects caused by actuator attacks and integrates the advantages of both data-driven modeling and model-based control design. Finally, several comparative case studies show the effectiveness of the proposed secure control scheme.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"436-446"},"PeriodicalIF":9.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142637322","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-13DOI: 10.1109/TCYB.2024.3488371
Meng Zhai;Tong Yang;Qingxiang Wu;Shudong Guo;Ruiping Pang;Ning Sun
Underactuated systems are a class of systems in which the number of control inputs is less than the degrees of freedom (DoFs) to be controlled. With the increasing demand for the control performance of underactuated systems, the current research on their optimization of steady-state performance is no longer sufficient. However, owing to limited control inputs, ensuring their transient performance is often difficult. Moreover, some specific composite variables in underactuated systems should be kept within the preset ranges, which poses a significant challenge to collision avoidance safety. In addition, the sensor noises are also an issue that cannot be ignored. To this end, an extended Kalman filtering-based nonlinear model predictive control method for underactuated systems is developed in this article. The key feature of this method is that it simultaneously ensures accurate positioning, multiple constraints, and obstacle avoidance. Specifically, by adding an artificial potential field as an obstacle avoidance penalty term in the cost function and dynamically assigning weight coefficients, efficient collision avoidance control is achieved. Furthermore, it is combined with the extended Kalman filtering and jointly applied to underactuated systems with sensor noises. To the best of our knowledge, it is the first control method that simultaneously considers full-state constraints, specific composite variable constraints, control input and its increment constraints, as well as obstacle avoidance in underactuated systems. The satisfactory control performance of the proposed method is validated by implementing it on two typical underactuated systems, that is, four-DoF overhead cranes and five-DoF tower cranes.
{"title":"Extended Kalman Filtering-Based Nonlinear Model Predictive Control for Underactuated Systems With Multiple Constraints and Obstacle Avoidance","authors":"Meng Zhai;Tong Yang;Qingxiang Wu;Shudong Guo;Ruiping Pang;Ning Sun","doi":"10.1109/TCYB.2024.3488371","DOIUrl":"10.1109/TCYB.2024.3488371","url":null,"abstract":"Underactuated systems are a class of systems in which the number of control inputs is less than the degrees of freedom (DoFs) to be controlled. With the increasing demand for the control performance of underactuated systems, the current research on their optimization of steady-state performance is no longer sufficient. However, owing to limited control inputs, ensuring their transient performance is often difficult. Moreover, some specific composite variables in underactuated systems should be kept within the preset ranges, which poses a significant challenge to collision avoidance safety. In addition, the sensor noises are also an issue that cannot be ignored. To this end, an extended Kalman filtering-based nonlinear model predictive control method for underactuated systems is developed in this article. The key feature of this method is that it simultaneously ensures accurate positioning, multiple constraints, and obstacle avoidance. Specifically, by adding an artificial potential field as an obstacle avoidance penalty term in the cost function and dynamically assigning weight coefficients, efficient collision avoidance control is achieved. Furthermore, it is combined with the extended Kalman filtering and jointly applied to underactuated systems with sensor noises. To the best of our knowledge, it is the first control method that simultaneously considers full-state constraints, specific composite variable constraints, control input and its increment constraints, as well as obstacle avoidance in underactuated systems. The satisfactory control performance of the proposed method is validated by implementing it on two typical underactuated systems, that is, four-DoF overhead cranes and five-DoF tower cranes.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"369-382"},"PeriodicalIF":9.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142610721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-13DOI: 10.1109/TCYB.2024.3489438
Xingchen Yang;Zongtian Yin;Yixuan Sheng;Dario Farina;Honghai Liu
As a primary effector of humans, the hand plays a crucial role in many aspects of daily life. Recognizing multidegree-of-freedom hand movements from muscle activity helps infer human motion intentions. Solving this problem has direct applications in prosthetic and exoskeleton control. Here, we propose a self-supervised learning algorithm inspired by muscle synergies to achieve simultaneous estimation of wrist rotation (supination/pronation) and hand grasp (open/close) from sonomyography—the muscle deformation detected by a wearable ultrasound array. Unlike conventional methods collecting both muscle activity and hand kinematics for supervised model calibration, this algorithm only uses unlabeled forearm ultrasound signals for self-supervised wrist and hand movement estimation, where movement labels are auto-generated. The performance of the proposed algorithm was experimentally evaluated with ten participants including an amputee. Offline analysis demonstrated that the proposed algorithm can accurately estimate simultaneous wrist rotation and hand grasp movements ( $r_{textrm {wrist}}$ and $r_{textrm {hand}}$ were 0.98 and 0.94 for the able-bodied, and 0.98 and 0.90 for the amputee, respectively). Notably, the performance of the self-supervised learning was superior to the supervised learning for the amputee. Online experiments demonstrated that intended wrist and hand movements can be deciphered in real time, enabling accurate control of a virtual hand. This study will open up a new avenue for the sonomyographic human-machine interaction.
{"title":"Self-Supervised Learning for Intuitive Control of Prosthetic Hand Movements via Sonomyography","authors":"Xingchen Yang;Zongtian Yin;Yixuan Sheng;Dario Farina;Honghai Liu","doi":"10.1109/TCYB.2024.3489438","DOIUrl":"10.1109/TCYB.2024.3489438","url":null,"abstract":"As a primary effector of humans, the hand plays a crucial role in many aspects of daily life. Recognizing multidegree-of-freedom hand movements from muscle activity helps infer human motion intentions. Solving this problem has direct applications in prosthetic and exoskeleton control. Here, we propose a self-supervised learning algorithm inspired by muscle synergies to achieve simultaneous estimation of wrist rotation (supination/pronation) and hand grasp (open/close) from sonomyography—the muscle deformation detected by a wearable ultrasound array. Unlike conventional methods collecting both muscle activity and hand kinematics for supervised model calibration, this algorithm only uses unlabeled forearm ultrasound signals for self-supervised wrist and hand movement estimation, where movement labels are auto-generated. The performance of the proposed algorithm was experimentally evaluated with ten participants including an amputee. Offline analysis demonstrated that the proposed algorithm can accurately estimate simultaneous wrist rotation and hand grasp movements (\u0000<inline-formula> <tex-math>$r_{textrm {wrist}}$ </tex-math></inline-formula>\u0000 and \u0000<inline-formula> <tex-math>$r_{textrm {hand}}$ </tex-math></inline-formula>\u0000 were 0.98 and 0.94 for the able-bodied, and 0.98 and 0.90 for the amputee, respectively). Notably, the performance of the self-supervised learning was superior to the supervised learning for the amputee. Online experiments demonstrated that intended wrist and hand movements can be deciphered in real time, enabling accurate control of a virtual hand. This study will open up a new avenue for the sonomyographic human-machine interaction.","PeriodicalId":13112,"journal":{"name":"IEEE Transactions on Cybernetics","volume":"55 1","pages":"409-420"},"PeriodicalIF":9.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142610701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：