Robotics and Autonomous Systems最新文献_第2页

Asymmetric bounded fuzzy adaptive control for uncertain coordinative multiple robot manipulators 不确定多机器人协调臂的非对称有界模糊自适应控制

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-26 DOI: 10.1016/j.robot.2025.105198

Yongqing Fan, Lin Yang, Zhen Li

{"title":"Asymmetric bounded fuzzy adaptive control for uncertain coordinative multiple robot manipulators","authors":"Yongqing Fan, Lin Yang, Zhen Li","doi":"10.1016/j.robot.2025.105198","DOIUrl":"10.1016/j.robot.2025.105198","url":null,"abstract":"<div><div>This work presents a fuzzy adaptive control technology for determining the desired trajectory of collaborative robot manipulators when grasping a general object. While the classical fuzzy logic systems (FLSs) are commonly used to compensate for some unknown nonlinear continuous functions, their approximation accuracies are often limited. To address this issue, a non-zero time-varying parameter is introduced in the input of Mamdani type or Takagi–Sugeno (T–S) type FLSs. This parameter allows for universal approximation, enabling the system to automatically adjust the approximation precision through adaptive rules. The unknown nonlinear continuous functions are represented using a combined form of homogeneous functions, which are then approximated using FLSs. Unlike previous fuzzy adaptive control schemes, this approach overcomes the limitation of a finite universal approximation domain. Additionally, the proposed method can calculate the coefficients of consequents in T–S type FLSs, reducing the computational load of the controller. The effectiveness of the proposed sliding mode surface is demonstrated in ensuring the required tracking performance, with all signals in the closed-loop system being uniformly ultimately bounded (UUB). The efficiency of the control scheme is further demonstrated through various simulation results.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105198"},"PeriodicalIF":5.2,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222439","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multi-Agent Reinforcement Learning for Zero-Shot Coverage Path Planning with Dynamic UAV Networks 基于多智能体强化学习的动态无人机网络零射击覆盖路径规划

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-26 DOI: 10.1016/j.robot.2025.105163

José P. Carvalho, A. Pedro Aguiar

{"title":"Multi-Agent Reinforcement Learning for Zero-Shot Coverage Path Planning with Dynamic UAV Networks","authors":"José P. Carvalho, A. Pedro Aguiar","doi":"10.1016/j.robot.2025.105163","DOIUrl":"10.1016/j.robot.2025.105163","url":null,"abstract":"<div><div>Recent advancements in autonomous systems have enabled the development of intelligent multi-robot systems for dynamic environments. Unmanned Aerial Vehicles play an important role in multi-robot applications such as precision agriculture, search-and-rescue, and wildfire monitoring, all of which rely on solving the coverage path planning problem. While Multi-Agent Coverage Path Planning approaches have shown potential, many existing methods lack the scalability and adaptability needed for diverse and dynamic scenarios. This paper presents a decentralized Multi-Agent Coverage Path Planning framework based on Multi-Agent Reinforcement Learning with parameter sharing and Centralized Training with Decentralized Execution. The framework incorporates a customized Rainbow Deep-Q Network, a size-invariant reward function, and a robustness and safety filter to ensure completeness and reliability in dynamic environments. Our training pipeline combines curriculum learning, domain randomization, and transfer learning, enabling the model to generalize to unseen scenarios. We demonstrate zero-shot generalization on scenarios with significantly larger maps, an increased number of obstacles, and a varying number of agents compared to what is seen during training. Furthermore, the models can also adapt to more structured maps and handle different tasks, such as search-and-rescue, without the need for retraining.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105163"},"PeriodicalIF":5.2,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Integrative AI framework for robotics: LLM-enabled reinforcement learning in object manipulation and task planning 机器人集成人工智能框架：llm在对象操作和任务规划中的强化学习

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-25 DOI: 10.1016/j.robot.2025.105197

Truong Nhut Huynh, Kim-Doang Nguyen

{"title":"Integrative AI framework for robotics: LLM-enabled reinforcement learning in object manipulation and task planning","authors":"Truong Nhut Huynh, Kim-Doang Nguyen","doi":"10.1016/j.robot.2025.105197","DOIUrl":"10.1016/j.robot.2025.105197","url":null,"abstract":"<div><div>The paper develops an innovative hybrid AI framework that combines contextual reasoning of a large language model (LLM) with adaptivity of reinforcement learning (RL) for improved robotic object manipulation and task execution. In particular, the proposed system integrates high-level task planning, where GPT-4 and an RL submodule collaboratively generate optimized task strategies, with low-level real-time control through RL, allowing for enhanced adaptability in dynamic environments. The experimental results demonstrate significant improvements in task success rates and operational efficiency compared to standalone RL and GPT-4 approaches. In static environments, the integrative approach achieved a 90% task success rate, with an average completion time of 42.1 s and only 1.1 retries, outperforming RL-only (72%) and GPT-4-only (78%) methods. In dynamic environments, our integrative system maintained an 85% success rate, compared to 65% for RL-only and 70% for GPT-4-only. For complex tasks, the hybrid model showed a substantial advantage, with an 80% success rate, highlighting its superior performance in tasks requiring both high-level reasoning and low-level precision control.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105197"},"PeriodicalIF":5.2,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159988","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A path planning decision system for unknown pipeline detection using UAVs 基于无人机的未知管道检测路径规划决策系统

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-24 DOI: 10.1016/j.robot.2025.105194

Dekai Lin, Ruitao Ma, Yin Zhao, Jiakuo Zhang, Shubin Liu, Hang Zhu

{"title":"A path planning decision system for unknown pipeline detection using UAVs","authors":"Dekai Lin, Ruitao Ma, Yin Zhao, Jiakuo Zhang, Shubin Liu, Hang Zhu","doi":"10.1016/j.robot.2025.105194","DOIUrl":"10.1016/j.robot.2025.105194","url":null,"abstract":"<div><div>This paper introduces the Pipeline-Aimed path Planner (PAPlanner), a sampling-based path planner for UAVs in unknown oil/gas pipelines. Its key contributions include a dynamic anchor point update strategy and path optimization that adapts to bends and diameter changes, eliminating redundant backtracking and enabling continuous exploration. By integrating real-time voxel maps, the algorithm optimizes paths to stay near the pipeline axis. Simulation results show that PAPlanner reduces average path length by 26.4% compared to the advanced MBPlanner method in the elbow scene experiment, demonstrating efficient safe trajectory maintenance. In the variable diameter scene experiment, where MBPlanner fails frequently, PAPlanner achieves a 0% failure rate. Real flight experiments validate its robustness with a 0.309 m average trajectory deviation from the axis, confirming reliable navigation. This work presents a novel framework enhancing UAV exploration efficiency in pipelines, overcoming limitations of existing algorithms for autonomous inspection in sensor-degraded confined environments.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105194"},"PeriodicalIF":5.2,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145222440","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CPP-DIP: Multi-objective coverage path planning for MAVs in dispersed and irregular plantations 分散和不规则人工林中MAVs的多目标覆盖路径规划

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-24 DOI: 10.1016/j.robot.2025.105193

Weijie Kuang, Hann Woei Ho, Ye Zhou

{"title":"CPP-DIP: Multi-objective coverage path planning for MAVs in dispersed and irregular plantations","authors":"Weijie Kuang, Hann Woei Ho, Ye Zhou","doi":"10.1016/j.robot.2025.105193","DOIUrl":"10.1016/j.robot.2025.105193","url":null,"abstract":"<div><div>Coverage Path Planning (CPP) is vital in precision agriculture to improve efficiency and resource utilization. In irregular and dispersed plantations, traditional grid-based CPP often causes redundant coverage over non-vegetated areas, leading to waste and pollution. To overcome these limitations, we propose CPP-DIP, a multi-objective CPP framework designed for Micro Air Vehicles (MAVs). The framework transforms the CPP task into a Traveling Salesman Problem (TSP) and optimizes flight paths by minimizing travel distance, turning angles, and intersection counts. Unlike conventional approaches, our method does not rely on GPS-based environmental modeling. Instead, it uses aerial imagery and a Histogram of Oriented Gradients (HOG)-based approach to detect trees and extract image coordinates. A density-aware waypoint strategy is applied: Kernel Density Estimation (KDE) is used to reduce redundant waypoints in dense regions, while a greedy algorithm ensures complete coverage in sparse areas. To verify the generality and scalability of the framework, TSP instances of varying sizes are solved using three methods: Greedy Heuristic Insertion (GHI), Ant Colony Optimization (ACO), and Monte Carlo Reinforcement Learning (MCRL). An object-based optimization is subsequently applied to further refine the paths. Additionally, CPP-DIP integrates ForaNav, our insect-inspired navigation method, for accurate tree localization and tracking. Experimental results show that MCRL provides a balanced solution, reducing travel distance by 16.9 % compared to ACO while maintaining comparable performance to GHI. It also improves path smoothness by reducing turning angles by 28.3 % and 59.9 % relative to ACO and GHI, respectively, and eliminates intersections. Computational resource comparisons further highlight that GHI scales efficiently with increasing waypoints, whereas ACO and MCRL incur higher computational costs. These results confirm the robustness, efficiency, and scalability of the proposed CPP-DIP.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105193"},"PeriodicalIF":5.2,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159986","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Behavior-based navigation of a two-wheeled self-balancing robot using a modified hybrid automaton 基于行为的改进混合自动机两轮自平衡机器人导航

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-19 DOI: 10.1016/j.robot.2025.105195

Mohsen Heydari Khalili, Majid Sadedel

{"title":"Behavior-based navigation of a two-wheeled self-balancing robot using a modified hybrid automaton","authors":"Mohsen Heydari Khalili, Majid Sadedel","doi":"10.1016/j.robot.2025.105195","DOIUrl":"10.1016/j.robot.2025.105195","url":null,"abstract":"<div><div>Due to advances in robotics science, mobile robots are being used in more and more applications worldwide, and the autonomous navigation of these robots is an important topic in their discussion. This paper focuses on the autonomous navigation of a two-wheeled self-balancing robot (TWSBR) in an unknown environment using behavior-based control in the form of a hybrid automaton. This hybrid automaton includes the behaviors “Go To Goal” and “Avoid Obstacle,” and to avoid the Zeno phenomenon between these two behaviors, another behavior is considered in between, called “Follow Wall,” which the robot uses to move around the obstacle. However, two bugs are identified in the conventional hybrid automaton. The first bug causes the robot to not follow the optimal path. Another bug is that the Zeno phenomenon occurs between the two behaviors “Follow Wall” and “Go To Goal,” causing odometry errors in the experimental environment. The results show that the modified hybrid automaton successfully corrects the bugs and works as intended. The navigation algorithm is designed for the point mass model, so it is transformed to the unicycle model using a transformation, which can be used as input to the TWSBR controller. After linearizing the dynamic equations of the robot around its equilibrium point, the pole placement method is used to create the TWSBR controller. By adding the Luenberger observer to estimate the state variables, the non-full-state feedback system is also controlled. The results of the simulations demonstrate that the whole system is functioning properly so that the robot follows the path determined by the navigation algorithm while maintaining its equilibrium.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105195"},"PeriodicalIF":5.2,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159987","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

EM-LSD: A lightweight and efficient model for multi-scale line segment detection EM-LSD：一种轻量级、高效的多尺度线段检测模型

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-17 DOI: 10.1016/j.robot.2025.105192

Shuo Hu, Liye Zhao, Qing Wang

{"title":"EM-LSD: A lightweight and efficient model for multi-scale line segment detection","authors":"Shuo Hu, Liye Zhao, Qing Wang","doi":"10.1016/j.robot.2025.105192","DOIUrl":"10.1016/j.robot.2025.105192","url":null,"abstract":"<div><div>To address the challenges of detecting line segments in dynamic and geometrically complex environments, EM-LSD, a lightweight and efficient line segment detection model, is introduced. Accurate and efficient detection of line segments is critical for tasks such as environmental modeling and localization in SLAM, where the failure to extract robust line features can result in unreliable mapping and trajectory estimation. The design of EM-LSD is guided by the limitations of existing methods: traditional approaches often fail to capture multi-scale and global features in noisy scenes, while deep learning models with multi-stage architectures impose high computational costs, making them unsuitable for real-time applications. Inspired by the observation that multi-scale feature extraction is essential for handling diverse geometric structures, EM-LSD incorporates a Dense Atrous Convolution (DAC) module to effectively capture multi-scale information with minimal computational overhead. Additionally, the need for robustness against structural complexities and noise led to the integration of dual decoders with a Channel-Spatial Multi-scale Attention (CSMA) module and a Multi-scale Atrous Deformable Block (MADB), enabling adaptive feature representation. Experimental results on the Wireframe and YorkUrban datasets validate EM-LSD’s superior accuracy, robustness, and real-time performance, emphasizing its capability to support resource-constrained SLAM applications. This model not only addresses the limitations of existing methods but also enhances the reliability of environment modeling and localization, offering inspiration for the development of lightweight and efficient detection frameworks.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105192"},"PeriodicalIF":5.2,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145120364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Neuromorphic visuotactile slip perception for robotic manipulation 机器人操作的神经形态视觉触觉滑动感知

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-16 DOI: 10.1016/j.robot.2025.105191

Yiming Qiao , Chaofan Zhang , Shaowei Cui , Lu Cao , Zhigang Wang , Peng Wang , Shuo Wang

{"title":"Neuromorphic visuotactile slip perception for robotic manipulation","authors":"Yiming Qiao , Chaofan Zhang , Shaowei Cui , Lu Cao , Zhigang Wang , Peng Wang , Shuo Wang","doi":"10.1016/j.robot.2025.105191","DOIUrl":"10.1016/j.robot.2025.105191","url":null,"abstract":"<div><div>Visuotactile sensing technology has received extensive attention in the tactile sensing community due to its stable high-resolution deformation sensing capabilities. However, the existing visuotactile sensing methods are far from humanoid neural information processing mechanism. To address this gap, we propose a neuromorphic visuotactile slip detection method named VT-SNN using Tactile Address-Event Representation (TAER) encoding combined with brain-inspired Spiking Neural Network (SNN) modeling in this paper. Our extensive experimental results demonstrate that the VT-SNN achieves slip detection accuracy of 99.59% and F1 score of 99.28%, which is comparable to Artificial Neural Networks (ANNs) while exhibiting significant advantages in power dissipation and inference time. Furthermore, we deployed the VT-SNN on Intel neuromorphic computing chip–Loihi and performed closed-loop slip-feedback robotic manipulation tasks such as bottle-cap tightening and loosening. Our closed-loop neuromorphic visuotactile sensing system shows significant promise for high accuracy, low latency, and low power dissipation for robotic dexterous manipulation.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105191"},"PeriodicalIF":5.2,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145098643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

RRT-based CPC: A configuration planning method for continuum robots using Rapidly-exploring Random Tree algorithm 基于快速探索随机树算法的连续体机器人构型规划方法

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-15 DOI: 10.1016/j.robot.2025.105190

Qiqi Pan , Hongbo Wang , Yongfei Feng , Shijie Guo , Jingjing Luo

{"title":"RRT-based CPC: A configuration planning method for continuum robots using Rapidly-exploring Random Tree algorithm","authors":"Qiqi Pan , Hongbo Wang , Yongfei Feng , Shijie Guo , Jingjing Luo","doi":"10.1016/j.robot.2025.105190","DOIUrl":"10.1016/j.robot.2025.105190","url":null,"abstract":"<div><div>Obstacle-aware configuration control represents a critical challenge in the deployment of continuum robots for advanced applications such as robotic-assisted laparoscopic surgery and intelligent industrial grasping systems. At present, in order to realize the obstacle avoidance function of flexible robots, inverse kinematic calculations are usually unavoidable. The problems of large amount of computation, long solution time, and non-convergence of results make the configuration control for flexible robots still challenging. Most of the current studies use the inverse kinematics calculation of end tracking, and for flexible robots with multiple degrees of freedom, the success rate of obstacle avoidance is low and the computational cost is large. In this paper, a three-segment continuum configuration planning method based on Rapidly-exploring Random Tree (RRT) algorithm is proposed, in which the rough obstacle avoidance path is obtained by RRT algorithm, then the three-segment fitting is carried out by using the second-order Bézier curve, and the length error is evaluated to meet the planning requirements. Experiments such as obstacle avoidance tests, the arrival of target endpoints at different positions and different obstacle environments show that the proposed method can effectively map the feasible solution to the actual configuration. Compared with the inverse kinematics method, the proposed approach improves the success rate of obtaining feasible solutions by at least 14.8% and reduces the solution time by at least 55%. In addition, no prior curvature information and traditional inverse kinematics calculation are needed for the configuration control.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"195 ","pages":"Article 105190"},"PeriodicalIF":5.2,"publicationDate":"2025-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145098644","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Robust dexterous hand control strategy cascading bare hand pose estimation and joint jitter suppression 鲁棒灵巧手控制策略级联徒手位姿估计与关节抖动抑制

IF 5.2 2区计算机科学

Robotics and Autonomous Systems Pub Date : 2025-09-11 DOI: 10.1016/j.robot.2025.105189

Mingqi Chen , Feng Shuang , Shaodong Li , Xi Liu

{"title":"Robust dexterous hand control strategy cascading bare hand pose estimation and joint jitter suppression","authors":"Mingqi Chen , Feng Shuang , Shaodong Li , Xi Liu","doi":"10.1016/j.robot.2025.105189","DOIUrl":"10.1016/j.robot.2025.105189","url":null,"abstract":"<div><div>Vision-based dexterous hand control via human hand intuition has great potential in improving control naturalness and immersion, which further achieves better dexterity and generalization. However, challenges still exist in robust control, which is affected by environmental issues including estimation fluctuations and human hand physiological tremor. Hand pose estimation suffers from self-occlusions and self-similarities, and problem exists in balancing stability and hysteresis when suppressing jitters. In this paper, we develop a novel dexterous hand control strategy cascading bare hand pose estimation and joint jitter suppression to enhance controlling robustness. The bare hand pose estimation network utilizes CNNs, ASCS-RL and a biologic-awared refinement module. CNNs extract hand pose features, ASCS-RL obtains accurate hand joint locations. A biological-awared refinement module considering joint movement coupling is novelly modeled and proposed to better refine global hand pose. Meanwhile, joint jitters are reanalyzed, which consist of physiological tremor and error fluctuation. A zero-delay low pass filter with threshold is then introduced to suppress joint jitters. Ablation studies validate the effectiveness of the proposed estimation modules. Best accuracy on ICVL is shown in comparative experiments with recent works to the best of our knowledge, with state-of-the-art accuracy also achieved on other two datasets. Dexterous hand control experiment is finally carried out, where joint jitters are effectively suppressed via the suppression algorithm, and robust dexterous hand control is achieved using the proposed strategy performing static gestures and dexterous object interactions.</div></div>","PeriodicalId":49592,"journal":{"name":"Robotics and Autonomous Systems","volume":"194 ","pages":"Article 105189"},"PeriodicalIF":5.2,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145095142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0