Efficient memory management is essential for the stability and long-term performance of mobile robots in Simultaneous Localization and Mapping (SLAM). However, existing methods often struggle to control redundancy in keyframes and map points, leading to reduced efficiency, increased latency, and potential system failure due to resource constraints. Achieving high accuracy in both mapping and trajectory estimation while maintaining a compact state representation remains a key challenge for scalable and efficient SLAM systems. To address this issue, this paper proposes an efficient long-term visual SLAM method based on sparse prior embedding and nonlinear score-guided sparsification for memory-constrained environments. The approach embeds keyframe information into sparse prior factors, avoiding global coupling while preserving system sparsity and consistency. Additionally, a nonlinear scoring function combining parallax and descriptor uniqueness is introduced to guide map point sparsification within the sliding window. This strategy enables efficient state graph management, achieving compact global map representations and effective observation constraints. The proposed method has been implemented in a complete visual SLAM system and evaluated through long-term real-world mapping experiments on an embedded robotic platform. Experimental results demonstrate that the approach significantly reduces memory consumption while maintaining trajectory and mapping accuracy. Furthermore, the method ensures real-time execution and deployment potential, indicating its suitability for large-scale SLAM tasks in resource-constrained and long-duration operational scenarios.

This paper addresses the challenge of robotic path planning in high-temperature operational environments. Aiming at the limitation of traditional methods that rely on geometric information while ignoring thermal radiation data, a potential field method path planning based on temperature-enhanced point cloud maps is proposed. By fusing the data of LiDAR and thermal imager, point cloud maps with temperature attributes are generated. Based on the traditional potential field method, the potential field function of thermal radiation repulsion was introduced. The temperature gradient of the heat source is transformed into the potential energy constraint of path planning to achieve path generation in high-temperature areas. The experimental results show that, compared with the traditional artificial potential field method, this study provides a new technical solution for the safe operation of robots in a thermal environment.

A methodology for impedance control based on mechanical vibration theory was presented to select stiffness and damping to meet the desired dynamic response using modal analysis. This methodology provided profound insights into how changes in stiffness and damping matrices influence a robot’s dynamic response under impedance control, by determining natural frequencies and damping ratios in the modal space. In this paper, our analytical and experimental results challenge the claim of a recent paper stating that the use of a null-space projection does not affect the outcomes, which appears to be a claim that is valid based only on their specific choice of stiffness and damping matrices and robot configuration. Through experiments using a 7-DOF Franka Emika Panda robot, we demonstrate the influence of both stiffness/damping matrices and the weighting matrix of null-space control on the responses of robot manipulators with degrees of redundancy performing primary and secondary tasks. We also introduce and show insights of the benefits of switching between the statically and the dynamically consistent projection matrices, contributing to the broader understanding of the dynamic response of redundant robotic manipulators.

Effective impact mitigation is a critical aspect of ensuring the stable motion of legged robots, especially in challenging environments. This work introduces a leg motion control system integrating nonlinear active compliance and impedance control with a balance controller. Designed for high-speed dynamic locomotion, the proposed system achieves superior impact absorption and precise leg position tracking without relying on passive spring mechanisms. The proposed nonlinear control, featuring state-dependent adaptive laws for stiffness and damping, outperforms linear alternatives in vibration suppression while maintaining tracking accuracy. A nonlinear PD parameter tuning method and optimized footstep locations are utilized for the balance controller to ensure quick recovery from external disturbances. Experimental validation on the SCIT Dog platform demonstrates exceptional cushioning capability during high-speed maneuvers and effective full-degree-of-freedom impact mitigation, including lateral disturbances. Crucially, this work advances beyond prior research by demonstrating active springless leg compliance/impedance control with formal stability guarantees and represents a fundamental locomotion layer that is essential for enabling future autonomous navigation capabilities in quadruped robots operating in unstructured environments.

To address the limitation of maximum jump height constrained by the maximum torque of hip joint motors in bipedal wheeled robots during trajectory planning using the dual-mass linear spring model (DMLSM), this study aims to reduce the maximum required hip joint torque for achieving equivalent jump heights. Concurrently, reducing the maximum rotational speed of hip joint demanded by equivalent jumps is essential to minimize motor wear and enhance system stability. Building upon the dual-mass model, we propose a trajectory planning method based on hip joint motor state optimization (TPM-HJMSO). This method integrates a nonlinear spring model, cubic polynomial interpolation, and a min–max optimization approach with nonlinear constraints. Simulation results demonstrate two critical advances: TPM-HJMSO achieves 163.16% higher jump heights than the DMLSM under identical maximum hip torque constraints, while reducing maximum hip motor speed by 37.17% when attaining equivalent jump heights. These outcomes validate the method’s superior performance. Precise trajectory tracking is demonstrated through feedforward-plus-PD control in both Webots simulations and physical prototype experiments, verifying TPM-HJMSO’s effectiveness.

In unstructured environments, Delta robots face challenges in achieving high vision-guided grasping precision due to dynamic lighting conditions and workpiece diversity. This paper designs an integrated solution that combines RGB-D multimodal learning with an enhanced Mask R-CNN framework. Initially, a dual-stream ResNet50-FPN backbone network is designed to achieve cross-modal adaptive alignment via hierarchical feature fusion. Subsequently, a depth-guided attention module is incorporated to bolster robustness against material ambiguity and reflective interference. Moreover, a dynamic depth estimation algorithm is employed to significantly improve target localization accuracy and stability. Finally, real-time trajectory tracking is realized by integrating PD control with Jacobian mapping. Experimental results validate the efficacy of the proposed method, offering an efficient and reliable approach for industrial robotic applications.

This paper addresses the challenge of balance control for the underactuated triple pendulum robot (UTPR) using a model-free reinforcement learning (RL) strategy. A curriculum-based Soft Actor-Critic strategy, with a quadratic form and an integral term in the reward function (CSAC-QI), is proposed. By incorporating the integral of cumulative joint angle errors into the reward function, the CSAC-QI method significantly reduces steady-state errors and enhances control precision. CSAC-QI improves convergence efficiency through an adaptive curriculum learning (CL) framework that enables a structured transition from simpler to more complex tasks. To enhance control robustness, motor friction identification and domain randomization are implemented during training, thereby equipping the UTPR to cope with real-world uncertainties. Simulation experiments demonstrate superior performance of the CSAC-QI method in handling larger initial joint deviations, achieving accurate end-effector positioning, and maintaining balance under dynamic randomization, sensor noise, and external disturbances. Notably, the trained policy is directly deployed on the UTPR prototype, where it successfully maintains balance in real-world conditions.

The spiral structure of interventional robots is widely utilized for vascular dredging procedures. Variations in the spiral structural parameters of these robots exert a significant impact on the hemodynamic parameters within the vessel during the interventional process; therefore, optimizing the design of these parameters constitutes a crucial prerequisite for ensuring safe diagnosis and treatment. Based on the bidirectional fluid-structure interaction (FSI) between pulsatile blood and elastic blood vessels, theoretical calculations and experimental measurements of the fluid velocity inside the vessel during robotic intervention were conducted. The results demonstrate that the trends and magnitudes of the calculated values are essentially consistent with those of the measured values. Furthermore, the range and variance methods in orthogonal design were applied to numerically investigate the effects of three key spiral parameters – spiral pitch, spiral groove depth, and thread profile – on hemodynamic and vascular parameters, including blood velocity, blood pressure, vascular wall shear stress, and vessel deformation. The findings indicate that spiral groove depth is the primary and most significant factor influencing blood velocity, blood pressure, and vessel deformation. In contrast, spiral pitch emerged as the main and critical factor affecting vascular wall shear stress. Meanwhile, thread profile is found to be a secondary factor with relatively minor impacts on all the aforementioned parameters. To achieve the objective of safe interventional diagnosis and treatment, it is recommended that the robot’s spiral structure be designed with a smaller spiral groove depth, a moderate spiral pitch, and a thread profile featuring strong cutting capability.

Legged robots operating in irregular environments are often subjected to compound disturbances such as tilts and vibrations, which can degrade attitude stability and motion reliability. This paper presents a real-time disturbance-adaptive control framework for a hexapod robot. The proposed system integrates quaternion-based attitude estimation using an extended Kalman filter (EKF), a double-threshold pose classifier, and a modular gait library and is implemented on an embedded controller with a 2 ms control-loop latency. Analytical verification and laboratory experiments demonstrate that the proposed control loop achieves uniform ultimate boundedness (UUB) under deterministic hybrid disturbances composed of controlled tilt and vibration, with a mean recovery time of 5.7 s. These results demonstrate that a lightweight rule-based controller can ensure reliable posture recovery within the experimentally validated laboratory scenarios, providing a foundation for future extensions to more complex environments. The main contributions of this work are (1) a disturbance-adaptive gait selection architecture for quasi-static stabilization, (2) a noise-robust EKF-based attitude estimation and double-threshold pose determination scheme, and (3) a concise Lyapunov-based stability analysis demonstrating UUB of the closed-loop system.

Under the backdrop of rapid advancements in renewable energy technologies and increasingly urgent demands for performance enhancement in solar power generation manipulator, this study proposes a three-revolute-prismatic-spherical (3-RPS) parallel manipulator featuring similar isosceles triangular moving and fixed platforms with revolute joint axes perpendicular to the angular bisectors of each vertex. Based on this manipulator, a novel dish solar concentrator was designed. Taking this concentrator as the research subject, kinematic analysis was conducted to derive the inverse kinematics model and overall Jacobian matrix of the manipulator. Three performance indexes were established: workspace, global motion/force transmission performance, and global stiffness performance. The Non-dominated Sorting Genetic Algorithm II genetic algorithm was applied for multi-objective optimization of the parallel manipulator, followed by TOPSIS decision-making based on the entropy weight method to select the optimal solution from the Pareto frontier. The distribution characteristics of these indexes within the achievable workspace before and after optimization were compared through performance atlas methodology. Static analyses were subsequently performed using finite element analysis to validate the manipulator’s performance. Results demonstrate that the workspace index and global performance stiffness index achieved varying degrees of improvement, while the global motion/force transmission performance exhibited a slight overall decline. Nevertheless, under specific operational conditions, the optimized manipulator shows enhanced performance compared to the original design while strictly satisfying all constraint requirements.

The period-$1$ gait is a stable, efficient, and predictable walking pattern for bipedal robots, which is crucial for achieving precise control, energy optimization, and maintaining balance in complex environments. This study investigates the control of chaotic and multi-period gaits in a bipedal robot with impulsive actuation, focusing on the stabilization of such gaits into period-1 gaits and the expansion of their stable parameter regions. Three main contributions are presented: (1) stabilizing chaotic gaits to period-$1$ gaits using the Ott–Grebogi–Yorke (OGY) and bifurcation control methods; (2) converting multi-period gaits (e.g., period-$2$ and period-$4$) into period-$1$ gaits via bifurcation control; and (3) extending the stable parameter range of period-$1$ gaits by optimizing pulse thrust parameters. The bifurcation control method, though energy-intensive, proves more reliable than OGY method in preventing falls. Analytical and numerical results demonstrate that adjusting pulse thrust parameter $C$ significantly expands the stability domain, enhancing robustness against disturbances and parameter variations. The bifurcation control method proposed in this study is implemented via a constant thrust increment and validated through numerical simulations, which makes it easy to be applied in controller design. It provides both a theoretical foundation and a practical control scheme for achieving stable and efficient gait design in bipedal robots with impulsive actuation.

Preference-based reinforcement learning (PbRL) significantly simplifies the design of reward functions in reinforcement learning (RL) tasks. However, because of the tasks’ complexity, intransitive preferences, and sensitivity to preference errors, PbRL requires substantial feedback to achieve the desired performance. This extensive reliance on expert input notably increases the burden on participants. We have developed a novel framework: Self-teacher-learning preference-based reinforcement learning (STL-PbRL). In the teacher-led (TL) module, agents learn more reliable reward prediction models (RM) through TL PbRL. In the self-learning (SL) module, agents utilizing the preference comparison approach for trajectory segments integrate sparse but critical, easily designed task-oriented information into the feedback process. The STL-PbRL framework incorporates the SL module to refine the RM initially generated by the TL module. We have demonstrated that this integration significantly enhances RM by enabling RM to converge toward an optimal reward model that effectively supports achieving a training policy that meets task objectives. This streamlined and efficient STL-PbRL framework enables a more accurate and efficient training process. Our experimental results confirm that the SL module seamlessly integrates with existing PbRL algorithms, significantly reducing the need for feedback and alleviating the impact of errors in preference indications. This efficiency and effectiveness highlight that STL-PbRL innovates, simplifies, and enhances the RL training process across various applications.

The ability of multirotor unmanned aerial vehicles (UAVs) to perform accurately in windy environments is crucial for extended use in outdoor applications. To design UAVs to operate in these environments, most studies have focused on static performance metrics such as thrust-to-weight ratio and endurance, without directly considering closed-loop control performance. This work develops a simplified metric that serves as a predictor for achievable disturbance rejection performance, enabling efficient UAV design selection without requiring full-scale nonlinear simulations. A reduced-order model is introduced to capture key aerodynamic and actuation characteristics, allowing for rapid evaluation of UAV configurations. The metric is validated against high-fidelity nonlinear simulations, demonstrating strong correlation with actual control performance. By bridging the gap between UAV structural optimization and closed-loop control behavior, this approach provides a practical tool for integrating disturbance rejection capabilities into UAV design processes. The practical utility of this metric is supported by experimental findings from related wind tunnel studies of fully-actuated UAVs, which demonstrate that actual disturbance rejection performance aligns with the trends predicted by the simplified correlation function.

An underactuated nonholonomic autonomous robot with a Dubins car kinematics navigates through a planar environment that hosts an unpredictably moving rigid formation of point-wise targets. Both the number of the targets and their geometric configuration are unknown. The robot measures the relative positions of the targets within a limited sensing range but has no access to their velocities or other kinematic parameters. The robot has to approach the formation, and to subsequently move so that first, all targets are surrounded by the robot’s path, and second, a predefined distance to the currently nearest target is maintained. However, a lower bound on the turning radius of the robot typically makes perfect achievement of the second objective impossible since different targets become the nearest to the robot in the course of its movement. So a scheme for on-the-fly synthesis of a feasible and somewhat optimal approximation of the ideal circumnavigation path is offered. Based on this development, the paper presents a computationally efficient control law that accomplishes the mission as much as possible. The effectiveness of the advocated approach is justified by mathematically rigorous proofs of nonlocal convergence and is confirmed by simulation tests and experiments with real robots.

Force/motion transmissibility serves as an essential performance metric in the kinematic analysis of parallel mechanisms. The initial step in evaluating force/motion transmissibility involves identifying the basis elements of twist and wrench spaces. In parallel mechanisms consisting of PRS, RPS, SPR, and PU limbs, the identification of the basis elements of a given limb is relatively straightforward, as the limb constraints can be evaluated as pure forces and torques. However, in limbs where the axes of the joints are offset, determining these basis elements becomes quite complex. Typically, the constraints of these limbs involve the interconversion between general screws and line vectors or couples. This paper focuses on a 3-PRRU parallel mechanism. Procedures for identifying the basis elements are investigated. Screw theory is used to elucidate the relationships between the variations in the basis elements and the configurations of the mechanism. Moreover, the influence of an offset between two revolute joint axes on the force/motion transmissibility is systematically analyzed. The results indicate that, compared to the case without offsets, the introduction of offsets enhances the force transmissibility of the mechanism.

Physical human–robot collaboration (pHRC) has gained significant traction in industrial settings due to its exceptional flexibility and adaptability. However, robots often face unexpected nonlinear and random impact disturbances during task execution, which pose risks such as human injury and robotic instability. To address these challenges, this paper proposes a novel control strategy based on improved shear-thickening fluid control (ISFC) for enhancing human–robot cooperation. Drawing inspiration from the unique properties of shear-thickening fluids, ISFC is designed to ensure stable collaboration between humans and robots while effectively mitigating the effects of random impacts. Furthermore, an ISFC-based framework is introduced for human/dual-arm robot collaboration. Comprehensive simulations and experimental evaluations are conducted to validate the effectiveness and advantages of the proposed approach. The experimental results show that compared with the traditional linear admittance control (L-AC) and SFC, ISFC not only has excellent impact resistance (peak velocity<0.1 m/s under 30 N impacts) but also overcomes the stability defects of SFC in low-speed operation, reducing the convergence time by 94.1% (from 8.29 s to 0.50 s) and residual displacement by 49.3% (from 13.21 mm to 6.7 mm). In dual-arm collaboration, the position synchronization error remains below 8 mm. These improvements provide a new idea for the application of human–robot collaboration.

This work describes the design and validation of a one-size-fits-all assistive exoskeleton for the upper limb that is self-adjustable to the wearer subject. The assistance function is performed even if the joints of the exoskeleton are not aligned with the joints of the subject; therefore, it does not require personalized adjustments during the wearing phase. The device is composed of a distal articulated system with a prismatic interface towards the body segment and a flexible proximal architecture that shifts the actuation towards the pelvis to limit the alteration of the subject’s center of gravity. In vivo experiments in the laboratory demonstrate the ability to alleviate muscular effort, and home-based experiments in performing daily activities show excellent perceived usability and acceptability in an elderly population. Furthermore, a proposed biomechanical model estimates the ability of the exoskeleton to contain the joint constraint reactions in the subject during the assistance phase.

Gravity compensation, widely employed in mechanisms such as manipulators and exoskeletons, serves to minimize unnecessary energy consumption. Enhancing the performance of gravity compensation mechanisms (GCMs) is therefore of significant practical importance. One promising approach is the development of novel design methodologies. Existing GCMs utilizing cable-pulley units exhibit limitations in terms of structural compactness and efficiency. To address these issues, this paper proposes an innovative X-type gravity compensation unit featuring an adjustable cable sequencing arrangement. The application of this unit to a passive waist exoskeleton (PWE) is further introduced. Given the potential influence of pulley size and positioning on the actual spring elongation, a pulley-angle error model is established to analyze the effect of varying pulley positions. Simulations confirm the validity of the model and demonstrate that pulley size is a non-negligible factor in system performance. Finally, an experimental platform is constructed to evaluate the performance of the PWE. Experimental results indicate a notable improvement in exoskeleton performance attributable to the unique design of the X-type unit.

Hemiplegia, the paralysis of one side of the body, is a common effect of stroke and provides unique challenges for afflicted individuals, including asymmetric body strength and limited mobility, especially in the sit-to-stand (STS) motion. Reducing weight-bearing asymmetry during STS is important for improving mobility outcomes of hemiplegic patients. To address this concern, a semi-wearable STS assistive robot is proposed to provide assistive force and motion guidance during the STS motion. It is a planar 2-DoF assistive robot attached near the hip, designed to reduce weight-bearing asymmetry and facilitate correct execution of the STS motion by guiding the user along a target STS path and constraining pelvic motion in the frontal plane, controlled using a single worn IMU. The method for generating unique target STS paths and assistive robot design is presented. Experiments on healthy test subjects with the motion of one leg constrained were conducted to determine the changes and correlations in force and motion parameters when using the assistive robot during STS. The assistive robot improved rising STS asymmetry in some test subjects and reduced stabilization weight-bearing asymmetry in all test subjects. Motion data showed that the assistive robot facilitated hip translation and tilt toward the test subjects’ constrained side, while a counter trunk tilt toward the unconstrained side was observed. The results of the experiments suggest that more active control of the hip position and tilt and providing real-time feedback during the STS motion could further improve the function of the robot.

Efficient global localization of mobile robots in symmetrical indoor environments remains a formidable challenge, given the inherent complexities arising from uniform structures and a dearth of distinctive features. This review paper conducts an in-depth investigation into the nuances of global localization strategies, focusing on symmetrical environments, such as extended corridors, symmetrical rooms, tunnels, and industrial warehouses. The study comprehensively reviews and categorizes key techniques employed in this context, encompassing probabilistic-based approaches, learning-based approaches, Simultaneous Localization and Mapping (SLAM)-based approaches, and optimization-based approaches. The primary goal is to provide a contemporary and thorough literature review, offering insights into existing global localization solutions, followed by extant methods tailored for symmetrical indoor spaces. Also, the paper addresses practical challenges associated with implementing various global localization techniques, contributing to a holistic understanding of their real-world applicability. Comparative experimental results demonstrate that hybrid approaches achieve superior localization accuracy in symmetrical environments compared to any single method alone. These experiments, conducted in indoor settings with different symmetry levels, highlight the hybrid approach’s robustness and precision in resolving symmetry-induced ambiguities. This work signifies a significant step forward in mobile robot global localization, which addresses symmetrical environments’ complexities by leveraging the strengths of hybrid methodologies.