本期机器人顶刊精选

Real-world robotic manipulation in homes and factories demands reliability, efficiency, and robustness that approach or surpass skilled human operators. We present a real-world reinforcement learning (RL) framework, RL-100, for achieving complete task success under a predefined evaluation protocol built on diffusion visuomotor policies. RL-100 unifies imitation and RL under a single clipped proximal policy optimization surrogate objective applied in the denoising process, yielding conservative, stable improvements across offline and online stages. To meet deployment latency, a lightweight consistency distillation compresses multistep diffusion into a one-step controller for high-frequency control. The framework is task, embodiment, and representation agnostic and supports both single-action and action-chunking control. We evaluated RL-100 on eight diverse real-robot tasks, from pushing and bowling to pouring, cloth folding, unscrewing, multistage juicing, and long-horizon box folding. Under our predefined protocol, RL-100 achieved 100% success in the evaluated trials (1000 of 1000 episodes), including up to 250 of 250 consecutive trials on one task. It matched or surpassed expert teleoperators in time to completion. Without retraining, a single policy attained ∼90% zero-shot success under environmental and dynamics shifts, adapted in a few-shot regime to substantial task variations (86.7%), and remained robust to human perturbations (about 96%). Our juicing robot served customers continuously for about 7 hours without failure when deployed zero-shot in a shopping mall. These results suggest a potential path to deployable robot learning by starting from human priors, aligning training objectives with human-grounded metrics, and reliably extending performance beyond human demonstrations.

To survive in dynamic and unstructured environments, animals have developed extraordinary multimodal locomotion capabilities. This biological insight drives the innovation of bioinspired multimodal robots, which are defined as robotic systems integrating and transitioning between two or more distinct modes of bioinspired locomotion. This Review explores the historical progression, key design considerations, and current challenges faced in the realm of multimodal robotics. We highlight recent advancements in robotic body design, including the integration of soft materials, innovative structure repurposing strategies, and the deployment of multirobot systems. These advancements facilitate seamless mode transitions through both active and passive structural reconfigurations. In the realm of path planning and motion control, the paradigm is progressively shifting from traditional graph-based approaches and discrete controllers to learning-based frameworks. To address the current void of standardized benchmarks in this field, we propose five performance metrics—number of modes, marginal cost of modality, component repurpose percentage, transition cost, and performance improvement—that provide a quantitative framework for evaluating multimodal robots in terms of design effectiveness and operational performance. Last, we delineate a strategic roadmap for the future that advocates for the integration of physical and computational intelligence in multimodal robotics, arguing that this convergence is crucial for enabling real-time behavioral adaptations to complex environmental stimuli and thereby enhancing the robustness and functionality of multimodal robotic systems.

Enabling quadrupedal robots to traverse complex terrains, from rugged outdoor environments to urban landscapes, requires seamless integration of multiple motor skills, smooth transitions between gaits, and high-speed perceptive locomotion using only onboard sensors. We present APT-RL (action pretrained transformer–based reinforcement learning), a unified framework that enables multiskill locomotion to achieve high-speed traversal in complex environments through autonomous skill transitions using only onboard perception and computation. Our approach generates large-scale, feature-rich two-dimensional (2D) motion datasets through trajectory optimization with simplified dynamics. These datasets enable training of diverse, reusable locomotion skills that transfer effectively to a real quadruped robot operating on complex uneven terrains. The resulting high-quality skills serve as strong priors for efficient learning of complex downstream tasks and extend naturally to 3D environments, enabling smooth, high-speed multiskill locomotion in deployed policy. Real-world experiments demonstrate the framework’s capabilities: The robot performed agile maneuvers through complex indoor obstacles and outdoor wild environments, including dynamic drop-down maneuvers that reached instantaneous peak speeds of up to 6 meters per second. A single onboard policy enabled robust traversal of diverse obstacles, including stairs, hurdles, stepping stones, gaps, and fallen branches, demonstrating the versatility and effectiveness of our approach.

The human brain can process spatial action through either egocentric (self-centered) or allocentric (environment-centered) perspectives. Despite this cognitive ability, teleoperation of robotic systems has been predominantly egocentric, wherein operators control a single robot from an inside-out perspective with limited situational overview. This paper proposes a scalable paradigm for simultaneously teleoperating multiple robotic systems from an allocentric perspective that enables a single operator to solve complex collaborative tasks. The perceptual information of all robots is fused into a joint virtual environment, which the operator can view from a top-down perspective to control the robots similar to puppets. We investigated differences between egocentric and allocentric multirobot teleoperation using both simulated and real robots. In the simulation study, study participants ( n = 15) reported a higher sense of embodiment over robot bodies when using egocentric control, enabling them to perform more precise manipulation after initial training. Using allocentric control, participants achieved speed increases of 84% in a parallelizable locomotion task and 50% in a situational overview task compared with egocentric control. Participants reported improved system usability and overview with allocentric control, and no significant differences in sense of agency were measured. These results were validated in a real-world study with two physical robots ( n = 6) and underscore the effectiveness of our allocentric teleoperation framework as a complementary approach to egocentric robotic teleoperation. Interfaces that support smooth switching between allocentric and egocentric control therefore enable operators to leverage the advantages of both paradigms depending on their specific task.

Conventional SLAM systems using visual or LiDAR data often struggle in poor lighting and severe weather. Although 4D radar is suited for such environments, its sparse and noisy point clouds hinder accurate odometry estimation, while the radar maps suffer from obscure and incomplete structures. Thus, we propose Super4DR, a 4D radar-centric framework for learning-based odometry estimation and gaussian-based map optimization. First, we design a cluster-aware odometry network that incorporates object-level cues from the clustered radar points for inter-frame matching, alongside a hierarchical self-supervision mechanism to overcome outliers through spatio-temporal consistency, knowledge transfer, and feature contrast. Second, we propose using 3D gaussians as an intermediate representation, coupled with a radar-specific growth strategy, selective separation, and multi-view regularization, to recover blurry map areas and those undetected based on image texture. Experiments show that Super4DR achieves a 67% performance gain over prior self-supervised methods, nearly matches supervised odometry, and narrows the map quality disparity with LiDAR while enabling multi-modal image rendering.

Real-world learning can push robot manipulation beyond imitation alone.

Work environments are often inadequate and lack inclusivity for individuals with upper-body disabilities. This paper presents a novel online framework for adaptive human-robot interaction (HRI) that accommodates users' arm mobility impairments, ultimately aiming to promote active work participation. Unlike traditional human-robot collaboration approaches that assume able-bodied users, our method integrates a mobility model for specific joint limitations into a hierarchical optimal controller. This allows the robot to generate reactive, mobility-aware behaviour online and guides the user's impaired limb to exploit residual functional mobility. The framework was tested in handover tasks involving different upper-limb mobility impairments (i.e., emulated elbow and shoulder arthritis, and wrist blockage), under both standing and seated configurations with task constraints using a mobile manipulator, and complemented by quantitative and qualitative comparisons with state-of-the-art ergonomic HRI approaches. Preliminary results indicated that the framework can personalise the interaction to fit within the user's impaired range of motion and encourage joint usage based on the severity of their functional limitations.

The Research Article “Observing a robot peer’s failures facilitates students’ classroom learning” by L. Chen et al. had a label error in Fig. 5A. The PF group was labeled 5.74, but the correct value is 6.12. The bar height, the error bar, and the standard error annotation (±0.30) for this group were all plotted correctly. The numerical label has been corrected, and the other data and conclusions are not affected.

Reliable robot autonomy hinges on decision-making systems that account for uncertainty without imposing overly conservative restrictions on the robot’s action space. We introduce Chance-Constrained Via-Point-Based Stochastic Trajectory Optimisation ( CC - VPSTO ), a real-time capable framework for generating task-efficient robot trajectories that satisfy constraints with high probability by formulating stochastic control as a chance-constrained optimisation problem. Since such problems are generally intractable, we propose a deterministic surrogate formulation based on Monte Carlo sampling, solved efficiently with gradient-free optimisation. To address bias in naïve sampling approaches, we quantify approximation error and introduce padding strategies to improve reliability. We focus on three challenges: (i) sample-efficient constraint approximation, (ii) conditions for surrogate solution validity, and (iii) online optimisation. Integrated into a receding-horizon MPC framework, CC-VPSTO enables reactive, task-efficient control under uncertainty, balancing constraint satisfaction and performance in a principled manner. The strengths of our approach lie in its generality, that is, no assumptions on the underlying uncertainty distribution, system dynamics, cost function, or the form of inequality constraints; and its applicability to online robot motion planning. We demonstrate the validity and efficiency of our approach in both simulation and on a Franka Emika robot. Videos and additional material are made available here: https://sites.google.com/oxfordrobotics.institute/cc-vpsto .

Underwater multi‐robot and swarm systems require advanced methodologies for controlling collective behavior. Conventional single‐robot techniques, such as tethering, sonar imaging, and acoustic localization, are not scalable or effective for swarm applications. This work introduces novel approaches for visual detection, relative localization, and control of a system of multiple autonomous underwater robots. In particular, we focus on agent‐to‐agent visual localization to inform control strategies and enable the emergence of coordinated swarm behaviors. The proposed methodologies have been implemented and tested on a fleet of autonomous underwater robots developed as part of the HSURF project. We present a thorough experimental validation of the integrated detection‐to‐control system, demonstrating its robustness and effectiveness in realistic, real‐world‐like conditions. This research marks a significant step forward in developing scalable solutions for underwater multi‐robot systems.

Accurate odometry estimation is fundamental to autonomous navigation, enabling precise localization and continuous motion tracking. While vision‐based and lidar‐based odometry systems have reached a high level of maturity, their performance deteriorates significantly in adverse environmental conditions, such as low visibility, inclement weather, or dynamic lighting. In contrast, 4D radar offers inherent robustness to such challenges, along with extended range and velocity‐sensing capabilities, making it a compelling alternative for long‐term, all‐weather perception. In this paper, we present an end‐to‐end neural network‐based odometry framework that directly leverages raw ADC data from 4D radars, surpassing conventional methods that are limited to processing only 4D point cloud information. We introduce a lossless encoding scheme that converts the raw ADC signals into structured heatmap representations, preserving essential spatial and temporal information. To extract meaningful features from this rich representation, we design a hybrid attention architecture that combines intra‐frame self‐attention for enhanced spatial understanding with inter‐frame cross‐attention to model fine‐grained temporal dependencies. Comprehensive evaluations on the Coloradar dataset show that our method significantly outperforms state‐of‐the‐art radar‐based odometry approaches based on traditional feature matching or deep learning. It achieves superior accuracy and robustness, highlighting its strong potential for real‐world deployment in autonomous navigation systems under challenging conditions. Our code is available at https://github.com/MoYuGit/Deep_Radar_Odometry .

Robotic haircutting is an emerging application of service robotics in which safety is paramount. Unlike conventional collaborative or domestic tasks, haircutting requires prolonged operation near the human head, one of the most vulnerable body regions, and often involves sharp or heated tools. These characteristics make safety central to both functional reliability and user acceptance. This paper reviews the safety landscape of haircutting robots by examining risks across the full task workflow and synthesizing lessons from related domains, including collaborative robotics, healthcare robotics, and autonomous systems. It proposes a three‐layer framework to organize existing approaches and illustrate how safety can be embedded from hardware design to control logic. In addition, a series of proof‐of‐concept demonstrations illustrates the practical implementation of the proposed three‐layer safety framework in representative haircutting scenarios. The paper further highlights opportunities and unresolved challenges in ensuring safe human‐robot interaction in haircutting contexts, providing guidance for future research and deployment.

To address the limitations of traditional unmanned aerial vehicles (UAVs) and wall‐climbing robots in bridge‐pier defect detection, this research proposes a circular‐climbing robot specifically designed for bridge piers. First, based on the geometric characteristics of bridge piers and their construction environment, a circular robot frame is designed, and its mechanical behavior is analyzed to determine the conditions for stable climbing. Second, based on attitude information collected by multiple gyroscopes, an attitude estimation method for the circular robot and a multi‐wheel cooperative motion control algorithm are developed. On this basis, a multi‐motor master–slave fuzzy proportional integral derivative (PID) controller is implemented on a field programmable gate array (FPGA) platform to perform fuzzy inference on the attitude information, generate real‐time control parameters, and adaptively adjust the speeds of the crawling mechanisms, thereby achieving coordinated motion control of the multi‐wheel system. Finally, experiments are conducted to verify the system's reliability in terms of climbing stability and motion control accuracy. A comparative analysis further shows that the proposed robot exhibits a more stable mechanical structure and greater adaptability than existing wall‐climbing robots, and offers advantages in contact pressure adjustment, load capacity, and scalability, enabling it to carry various bridge pier inspection devices for different application needs. This work provides a practical platform for the automated health monitoring of bridge piers.

With the expansion of urban water supply networks and the increasing visibility of aging infrastructure, traditional manual operations and maintenance can no longer meet the demands for high precision, low disturbance, and trenchless intelligent solutions. Robotic technology is currently one of the most advanced technologies for the operation and maintenance of water supply pipeline networks, and it is steadily gaining popularity in industry research. This paper provides a systematic review of the current research status of intelligent operation and maintenance robots across three major functional areas: detection, cleaning, and repair. It describes in detail the technical principles, application advantages, and current constraints of different kinds of robots. Multi‐source sensors are used by detection robots to locate and identify defects. Cleaning robots remove dirt using mechanical, water jet, and ultrasonic technologies. Spray‐in‐place lining and UV curing techniques are used by repair robots to accomplish trenchless rehabilitation. This review contributes a functional‐performance matrix that reveals a critical gap in robots capable of both high‐traction cleaning and precise inspection, pointing to cable‐free, autonomous navigation, intelligent control, modular and reconfigurable designs as the most urgent research frontier. Provides a clear future development direction for intelligent operation and maintenance robots.

In recent years, the demand for autonomous navigation of underwater biomimetic robots in complex environments has continued to grow, among which monocular visual simultaneous localization and mapping (SLAM) has emerged as a key technology for achieving lightweight navigation and localization. However, typical underwater illumination conditions, such as low‐light and artificial lighting, significantly degrade image quality and feature observability, thereby severely limiting the robustness and accuracy of existing visual SLAM systems. To address these challenges, this paper proposes a monocular visual SLAM system for bionic robotic fish, termed GSL‐SLAM , built upon the ORB‐SLAM3 framework. The proposed system integrates adaptive underwater image enhancement with learning‐based feature extraction and matching to improve visual usability and feature matching stability under complex illumination conditions. A multi‐threaded system architecture is further adopted to ensure real‐time performance. Extensive experiments conducted on both a laboratory pool dataset and the public AquaticVision dataset demonstrate that the proposed method consistently outperforms ORB‐SLAM3 under various underwater lighting conditions, including natural light, low light, and artificial illumination. Compared with the baseline method, GSL‐SLAM achieves significant improvements in initialization efficiency, map completeness, and localization accuracy, with particularly notable advantages under low‐light and artificial lighting scenarios. Moreover, the system operates stably at approximately 20 Hz on the embedded platform, satisfying the real‐time perception and navigation requirements of the bionic robotic fish. The experimental results indicate that GSL‐SLAM substantially enhances the robustness and practicality of monocular visual SLAM in complex underwater lighting environments, providing an effective visual localization and mapping solution for lightweight autonomous underwater robots.

Autonomous multi-finger grasping is a fundamental capability in robotic manipulation. Optimization-based approaches show strong performance, but tend to be sensitive to initialization and are potentially time-consuming. As an alternative, the generator-evaluator-planner framework has been proposed. A generator generates grasp candidates, an evaluator ranks the proposed grasps, and a motion planner plans a trajectory to the highest-ranked grasp. If the planner doesn't find a trajectory, a new trajectory optimization is started with the next-best grasp as the target and so on. However, executing lower-ranked grasps means a lower chance of grasp success, and multiple trajectory optimizations are time-consuming. Alternatively, relaxing the threshold for motion planning accuracy allows for easier computation of a successful trajectory but implies lower accuracy in estimating grasp success likelihood. It's a lose-lose proposition: either spend more time finding a successful trajectory or have a worse estimate of grasp success. We propose a framework that plans trajectories to a set of generated grasp targets, the evaluator estimates the grasp success likelihood at the terminal configuration of the planned trajectories, and the robot executes the trajectory most likely to succeed. Our experiments show our approach improves over the traditional generator-evaluator-planner framework across different objects, generators, and motion planners, and successfully generalizes to novel environments in the real world, including different shelves and table heights. Project Website: https://martinmatak.github.io/fpte/

Traditional sampling methodologies that depend on the manual operation of large submersibles are characterized by significant time consumption and labor intensity. The incorporation of autonomous underwater robotic arms within unmanned underwater vehicles (UUVs) presents a promising opportunity to markedly improve sampling efficiency. This paper delineates the design of a lightweight underwater robotic arm specifically engineered for the sampling of hydrothermal fluids at the seafloor. We establish a comprehensive modeling framework that integrates kinematic, rigid‐body, and hydrodynamic parameters. To address environmental uncertainties and unmodeled dynamics, we employ a long short‐term memory (LSTM)–augmented model in conjunction with a momentum‐based observer for end‐effector force estimation, obviating the need for external force/torque sensors. Furthermore, we propose a multi‐sensor fusion control strategy that synergizes visual perception with temperature tracking and force feedback, thereby facilitating autonomous and precise sampling of hydrothermal vents. The proposed methodology has undergone experimental validation, yielding promising outcomes.

Visibility underwater is challenging and degrades as the distance between the subject and the camera increases. That is why forward‐looking underwater computer vision tasks are difficult. We have collected underwater forward‐looking stereovision and visual–inertial image sets using two underwater imaging platforms, a stereo camera rig, and an ROV in the Mediterranean and Red Seas. To our knowledge, there are no other public data sets in the underwater environment with this forward‐looking camera‐sensor orientation that have published ground‐truth depth maps as well as pose. These data sets are critical for the development of several underwater applications, including autonomous obstacle avoidance, visual odometry, 3D tracking, Simultaneous Localization and Mapping and depth estimation through deep learning. The stereo data sets contain synchronized stereo images, and the visual–inertial data sets include monocular images and inertial measurement unit (IMU) measurements with millisecond‐level timestamp alignment. All data was collected in dynamic underwater environments with objects of known size. Both sensor configurations allow for scale estimation, with the calibrated baseline in the stereo setup and the IMU in the visual–inertial setup. Ground‐truth depth maps were created offline for both data set types using a commercial photogrammetry software (Agisoft Metashape). The ground truth is validated with multiple known measurements placed throughout the imaged environment. There are four stereo and 12 visual–inertial data sets in total, each containing thousands of images, with a range of different underwater visibility and ambient light conditions, natural and man‐made structures, and dynamic camera motions. The forward‐looking orientation of the camera plus the corresponding ground truth makes these data sets unique and ideal for testing underwater obstacle‐avoidance algorithms and for navigation close to the seafloor in dynamic environments. We show results from an experiment with a monocular depth estimation algorithm to demonstrate the applicability of the data sets. With our data sets, we hope to encourage the advancement of autonomous functionality for underwater vehicles in dynamic and/or shallow‐water environments.

The mechanical complexity of soft robots creates significant challenges for their model-based control. Specifically, linear data-driven models have struggled to control soft robots on complex, spatially extended paths that explore regions with significant nonlinear behavior. To account for these nonlinearities, we develop here a model-predictive control strategy based on the recent theory of adiabatic spectral submanifolds (aSSMs). This theory is applicable because the internal vibrations of heavily overdamped robots decay at a speed that is much faster than the desired speed of the robot along its intended path. In that case, low-dimensional attracting invariant manifolds (aSSMs) emanate from the path and carry the dominant dynamics of the robot. Aided by this recent theory, we devise an aSSM-based model-predictive control scheme purely from data. We demonstrate the effectiveness of our data-driven model in tracking dynamic trajectories across diverse tasks. We validate on high-fidelity, high-dimensional finite-element models of a soft trunk robot and Cosserat-rod-based elastic soft arms, with additional experiments confirming robust performance even in the presence of experimental noise. Notably, we find that five- or six-dimensional aSSM-reduced models outperform the tracking performance of other data-driven modeling methods by a factor up to 10 across all closed-loop control tasks.

Conventional field phenotyping methods are typically manual, time‐consuming, and destructive, creating a bottleneck for breeding progress. To address this challenge, robotics and automation technologies offer efficient sensing tools to monitor field evolution and crop development throughout the season. This study presents an end‐to‐end automated pipeline for terrestrial laser scanning (TLS) in plant breeding trials, built around a Husky ground robot equipped with a high‐resolution survey‐grade FARO 3D LiDAR scanner. Unlike prior TLS phenotyping approaches relying on manual scan placement or heuristic site selection, our system integrates a novel analytical 3D ray‐casting method for optimized TLS site planning with an offline route optimization algorithm that accounts for crop growth stages and field accessibility constraints. This enables efficient planning in complex breeding environments, reduces manual labor, and improves data collection efficiency, addressing scalability challenges in large breeding trials. Leveraging Real Time Kinematic‐Global Navigation Satellite System (RTK‐GNSS) and sensor fusion, the system achieved average errors below 0.6 cm for position and for heading, enabling point cloud registration with mean errors around 2 cm, comparable to traditional manual methods that require artificial targets. The platform was successfully deployed and evaluated in two distinct cotton breeding field layouts, demonstrating the platform's capability to autonomously collect accurate TLS data for quantitative plant phenotyping across varying plot configurations. The proposed autonomous phenotyping system advances scalable, efficient phenotyping workflows to support breeding programs for crop improvement, highlighting the potential for broader deployment in field phenomics.

Offshore wind turbine blade repair requires stable material removal and precise force regulation under curved‐surface contact and environmental disturbance. To address these challenges, this paper proposes an integrated constant‐force grinding method that combines a passive compliant end‐effector, a particle swarm optimization‐backpropagation neural network (PSO‐BP), and an improved active disturbance rejection control (ADRC) strategy. First, a passive compliant end‐effector with variable stiffness is designed to improve contact adaptability and reduce grinding impact on curved blade surfaces. Second, a PSO‐BP model is established to predict the material removal rate (MRR) and surface roughness (Ra) under different grinding conditions, thereby providing data‐driven support for process‐state evaluation and parameter scheduling. Third, based on a controller‐oriented force‐dynamics model, an improved ADRC framework integrating a tracking differentiator, nonlinear extended state observer, nonlinear state error feedback, and PSO‐BP‐assisted gain scheduling is developed for constant‐force grinding. A Lyapunov‐based analysis shows that the closed‐loop system is uniformly ultimately bounded under bounded disturbances and bounded scheduling error. Simulation and experimental results demonstrate that, compared with PID and standard ADRC, the proposed method achieves higher force‐tracking accuracy, stronger disturbance rejection, and better grinding quality. Under equivalent initial damage conditions, it produces the lowest post‐grinding surface roughness, indicating that the proposed method provides an effective solution for offshore blade grinding repair.

With the advancement of deep‐sea resource exploration and scientific research, underwater operations are evolving from rough manipulation to precise and intelligent tasks, imposing higher demands on the dexterity and environmental adaptability of underwater grippers. However, existing reviews often focus on superficial classifications based on actuation modes or application scenarios, there by lacking in‐depth analysis grounded in robotic grasping theory and system‐level integration. This paper presents a comprehensive review of underwater dexterous grippers, proposing a novel classification framework based on the integration of force/form closure theory and energy input methods. Under this framework, we systematically analyze the structural designs and mechanisms of grippers ranging from Fully Passive to Fully Active configurations. Furthermore, the review delves into kinematic and dynamic modeling methodologies for rigid, continuum, and rigid‐continuum coupled structures, alongside the challenges and solutions for underwater tactile and deformation perception systems. We also examine control strategies, progressing from low‐level open‐loop control to model‐based intelligent approaches, including the dynamic coupling within vehicle‐manipulator‐gripper systems (UVMS). To address the lack of standardized benchmarking, this study establishes a unified evaluation system incorporating quantitative metrics (e.g., Load‐to‐Weight Ratio, Normalized Durability Index) and benchmark protocols, providing a multi‐dimensional comparative analysis of existing technologies. Finally, the paper identifies emerging trends, suggesting that bio‐inspired rigid‐soft hybrid designs, multimodal perception fusion, and hybrid modeling approaches combining Physics‐Informed Neural Networks with Computational Fluid Dynamics (PINN‐CFD) represent critical directions for achieving highly reliable and dexterous underwater manipulation.

Accurate and robust localization is a fundamental requirement for service and inspection robots, particularly in feature-sparse indoor environments where traditional systems struggle due to a lack of distinct landmarks. While prior maps can enhance robustness, precise and compact maps capturing real-world details are often unavailable for new or frequently changing environments. This paper presents BIM-Loc, a novel discrepancy-aware LiDAR-based localization method that directly integrates Building Information Models (BIM) from the design phase. BIM-Loc simultaneously estimates trajectories aligned with the BIM coordinate system and identifies discrepancies between real-world observations and the as-designed BIM in an online fashion. Our core contributions include: (1) a novel multi-hit ray casting strategy for efficient BIM-point data association and projection of 3D observations into 2D texture space; (2) a pose graph optimization framework with BIM-integrated factors that enforces consistency among odometry, sequential scans, and BIM structures; and (3) a hierarchical Bayesian inference module that incrementally updates a continuous 2D surface representation for discrepancy detection, propagating updates from the pixel to the structure level. Extensive evaluations in both simulation and real-world applications demonstrate that BIM-Loc significantly outperforms state-of-the-art map-based methods in localization accuracy and robustness. More experimental results are available at our project website: https://bim-loc.github.io/bim-loc .

Large‐scale heavy‐duty hexapod robots have complex structures and driving methods, and require high real‐time motion control. Due to the large computational complexity of traditional dynamics control methods, which require the application of the robot's dynamic equations, it is difficult to achieve real‐time control. Moreover, the dynamics model of large‐scale heavy‐duty hexapod robots is difficult to accurately obtain because of the structural elastic deformation, making practical application quite challenging. This article proposed a Spring‐Muscle model to simplify the dynamics model of large‐scale heavy‐duty hexapod robots with symmetrical configuration and established a force‐controlled compliant control architecture for heavy‐duty hexapod robots. First, inspired by the EPH, the dynamics model of robot legs was simplified as a Spring‐Muscle model, and we analyzed dynamic equation of large‐scale heavy‐duty hexapod robots. Second, based on deformation force coordination constraints and tangential force distribution constraints, force distribution method is proposed to establish the overall dynamics balance relationship of the robot during motion. Finally, an overall control architecture is proposed and analyzed, which achieves smooth adjustment of the robot's foot‐ground interaction process, preventing large foot‐ground contact impacts. We validate our theory through a prototype of the principle. The related findings could improve the computational efficiency of robots and enhance their motion stability.

The rising levels of water pollution worldwide, particularly in urban areas due to industrial advancements, have created an urgent need for sustainable and intelligent surface water cleaning solutions. Robotic systems have emerged as one of the promising approaches to tackle floating debris and other contaminants that affect aquatic ecosystems. This review critically examines the recent advancements in surface water cleaning robots, focusing on their structural designs, operational principles, autonomy levels, and energy efficiency. Particular emphasis is placed on the integration of guidance, navigation, and control (GNC) systems that enable semi‐autonomous and fully autonomous functionality. Existing robotic platforms vary widely in their mechanical setups and cleaning mechanisms, ranging from simple net‐based collectors to more advanced systems equipped with visual sensors and embedded control logic. However, a majority of these systems still depend on limited sensing and require periodic human intervention. The lack of advanced imaging technologies for debris detection, real‐time classification, obstacle detection and navigation restrict their autonomous potential. In addition to technical comparisons, the paper highlights the limitations and research challenges, including power management, visual processing capabilities, and adaptive path planning. The need for more intelligent, low‐power, and modular systems is emphasized, particularly those capable of handling diverse aquatic environments with minimal supervision. By consolidating current knowledge and identifying performance gaps, this review aims to guide future research efforts in the development of efficient, scalable, and environmentally sustainable robotic systems for surface water cleaning. These advancements are essential in promoting long‐term solution for water pollution, water resource management, and supporting efforts to maintain cleaner aquatic ecosystems.

Global localization, which aims to obtain the absolute position of a planetary rover in georeferenced satellite images, is a challenging task in deep space exploration. The success of global localization is beneficial for long‐range exploration and therefore supports the emergence of significant scientific discoveries. Traditional high‐precision global localization methods usually generate the digital orthophoto map (DOM) product from the rover images in a certain place and obtain the rover position by registering the rover DOM product with the orbiter satellite images. This operation is commonly accomplished offline after the rover images are transmitted to the ground stations, due to the limited computational resources of the rover's online system. However, this procedure is time‐consuming and therefore impedes the operational efficiency of rovers. Moreover, the global localization operation is not always available owing to the difficulties in rover‐satellite image matching for extraterrestrial surfaces with poor textures. In this paper, an active global localization framework of planetary rover guided by the interest region perception is proposed, which is distinctly different from the commonly used passive mode. The main novelty of the proposed method is that the global localization module will only be activated automatically when recognizing certain interest regions which are helpful for the rover‐satellite image matching tasks. In this way, the efficiency of the global localization procedure can be improved greatly and the success rate can also be ensured. The described framework has been validated with the camera images from China's first Mars rover (Zhurong), and it is observed that the run time of the global localization operation can be reduced by 70% compared with the traditional processing average, indicating its potential for online application in global localization of rovers.

In unstructured environments, background regions and target objects often share visual attributes, such as texture, color, and structure. This similarity easily leads to feature confusion, posing significant challenges for grasp pose estimation. To address this issue, we propose BF‐GNet (RGB‐D fusion network for grasp pose estimation in complex background environments), an RGB‐D fusion network designed for grasp detection under complex backgrounds. At its core lies a scene‐aware Bimodal Foreground Enhancement Module, which employs a multilayer perceptron–based controller to adaptively weight RGB and depth modalities according to the input scene: prioritizing depth information in texture‐deficient conditions and relying more on RGB semantic cues when geometry is ambiguous. This adaptive fusion strategy effectively emphasizes salient foreground grasp regions while suppressing background interference. In addition, BF‐GNet integrates a Feature‐Channel‐Sharing 2D Selective Scan (FCSS) Module based on State Space Models, enabling long‐range dependency modeling with linear complexity. Experimental results show that BF‐GNet achieves accuracies of 98.9% and 94.8% on the Cornell and Jacquard data sets, and 95.12% and 91.8% on the cluttered CBRGD and large‐scale GraspNet‐1Billion data sets, respectively. Meanwhile, its computational cost remains as low as 0.72 GFLOPs, demonstrating strong competitiveness against existing methods. Furthermore, real‐world robotic experiments yield an average grasp success rate of 90.50%, validating the robustness, generalization capability, and practical value of BF‐GNet in complex scenarios. The video is available at https://youtube.com/shorts/eyLFwb3K-UE?si=B5U9xzXCt5YliGvB .

Robust localization of autonomous underwater vehicles is conventionally performed using dead‐reckoning, but its accumulated error can be excessive for long‐term or pioneering exploration when the Doppler Velocity Log cannot maintain bottom‐lock and when positioning references such as Ultra‐Short‐Base‐Line systems are unavailable. In this letter, we present an online localization that identifies current disturbances and corrects underwater position in real time without global references. An online Pruned Exact Linear Time method for optimal detection of statistical anomalies, such as current disturbances, is further expanded for real‐time dataset update and integrated with Kalman filter state estimation via recursive iteration. We demonstrate high accuracy in random‐current simulation and validate localization performance through implementation using 3‐h field data at two offshore sites. Furthermore, the comparative analysis indicates that, unlike other methods, online localization can effectively estimate underwater position in the presence of varying or stronger currents without dynamic model constraints.

The pretraining-finetuning paradigm has facilitated numerous transformative advancements in artificial intelligence research in recent years. However, in the domain of reinforcement learning (RL) for robot locomotion, individual skills are often learned from scratch despite the high likelihood that some generalizable knowledge is shared across all task-specific policies belonging to the same robot embodiment. This work aims to define a paradigm for pretraining neural network models that encapsulate such knowledge and can subsequently serve as a basis for warm-starting the RL process in classic actor-critic algorithms, such as Proximal Policy Optimization (PPO). We begin with a task-agnostic exploration-based data collection algorithm to gather diverse, dynamic transition data, which is then used to train a Proprioceptive Inverse Dynamics Model (PIDM) through supervised learning. The pretrained weights are then loaded into both the actor and critic networks to warm-start the policy optimization of actual tasks. We systematically validated our proposed method with 9 distinct robot locomotion RL environments comprising 3 different robot embodiments, showing significant benefits of this initialization strategy. Our proposed approach on average improves sample efficiency by 36.2% and task performance by 4.3% compared to random initialization. We further present key ablation studies and empirical analyses that shed light on the mechanisms behind the effectiveness of this method.

This article studies the time-optimal path planning problem for a convexified Reeds-Shepp (CRS) vehicle on a unit sphere, capable of both forward and backward motion, with speed bounded in magnitude by 1 and turning rate bounded in magnitude by a given constant. For the case in which the turning-rate bound is at least 1, using Pontryagin's Maximum Principle and a phase-portrait analysis, we show that the optimal path connecting a given initial configuration to a desired terminal configuration consists of at most six segments drawn from three motion primitives: tight turns, great circular arcs, and turn-in-place motions. A complete classification yields a finite sufficient list of 23 optimal path types with closed-form segment angles derived. The complementary case in which the turning-rate bound is less than 1 is addressed via an equivalent reformulation. The proposed formulation is applicable to underactuated satellite attitude control, spherical rolling robots, and mobile robots operating on spherical or gently curved surfaces. The source code for solving the time-optimal path problem and visualization is publicly available at https://github.com/sixuli97/Optimal-Spherical-Convexified-Reeds-Shepp-Paths.

Many hazardous tasks, such as those related to search and rescue scenarios, involve interactions with granular materials. In such examples, tactile sensing can be invaluable, since objects of interest are often not visible but may still be haptically salient. When a tool, such as a robot finger, approaches a buried object, the displaced granular material ahead of the tool begins to compact against the buried object. The resulting granular media jamming causes an increase in contact force felt on the robot finger without the finger making direct contact with the buried object. In this work, we present a framework for tactile‐based proximity sensing that leverages theoretical models of soil failure during soil‐tool interactions and empirical models of granular material failure zones to haptically detect the onset of granular media jamming and estimate the distance from a sensorized robot fingertip to an object buried in granular material. We investigated the effects of granular material type on proximity sensing accuracy by studying glass beads, poppy seeds, and playground sand. For various fingertip and object depths, we estimated the distance from a sensorized finger to a buried object during haptic exploration in granular materials on a real robot. We developed particle filters to estimate the force state on the robot fingertip in real‐time, and showed how occupancy maps can be continuously updated during haptic exploration within granular materials with a real robot. Our modeling framework can also be leveraged to simulate haptic interactions with granular materials prior to conducting real robot experiments.

With the increasing demand for inspection, maintenance, and operation in vertical and complex‐geometry environments, wall‐climbing robots have become essential equipment for high‐rise infrastructure maintenance, large‐scale industrial inspection, and emergency response. The codesign of adsorption and locomotion is fundamental to ensuring reliable attachment, stable mobility, and robust environmental adaptability. This paper provides a comprehensive review of major adsorption mechanisms—negative pressure, magnetic adsorption, electrostatic adsorption, bioinspired dry adsorption, and mechanical interlocking—and examines their underlying dynamics. Typical locomotion architectures, including wheeled, tracked, legged, peristaltic, and flipping mechanisms, are compared in terms of mobility characteristics and application boundaries across different surface conditions. Key engineering advances are summarized, including adsorption–locomotion coupling, curvature‐conforming structures, tunable adsorption, lightweight design, multimodal integration, and intelligent control. Considering emerging application scenarios, such as offshore wind turbines and nuclear power facilities—where resistance to wind loads, radiation, corrosion, and cross‐medium operation are required—future research directions are identified, such as multimodal adsorption fusion, compliant structural materials, enhanced environmental perception, and adaptive motion planning. This review aims to provide coherent technical guidance for the development of next‐generation high‐reliability, high‐adaptability wall‐climbing robotic systems

Rebar node tying on construction sites currently relies heavily on manual labor. The few existing mobile robots developed for this task are hindered by low end‐effector efficiency and poor motion continuity, characterized by redundant movements and travel paths. To address these issues, this article proposes a trajectory collaborative optimization strategy for the end‐effector of a rebar tying robot. The strategy optimizes the operational sequence of the end‐effector's horizontal, rotational, and descending modules, shortens their travel trajectories, and enhances the motion coupling between them. By establishing a pipeline parallel anticipation mechanism based on characteristic height layers, the strategy transforms the discrete “start‐stop” operational paradigm into a synchronized, continuous coupling framework. The performance of the proposed strategy was validated through comparative virtual simulations and physical experiments against an industrial‐standard Greedy Nearest Neighbor (GNN) baseline. The results demonstrate that for a standard eight‐node tying task, the strategy reduces the cumulative rotation angle by 46.2% and operational movements by 61.1%, while the total travel path is shortened by 43.3%. Most notably, the core system time is compressed by 38.0%, leading to a 136.5% improvement in overall operational efficiency compared with the non‐coupled baseline. The study further identifies a divergent performance gap, proving that the strategy's synergistic advantages—driven by the motion‐masking effect—become increasingly prominent in sparse task distributions. This research provides a robust, feasible solution for multi‐target sequencing and trajectory optimization in unstructured, large‐scale construction environments.

This paper presents a framework to prompt multi-robot teams with high-level tasks using natural language expressions. Our objective is to use the reasoning capabilities of language models in understanding and decomposing multi-robot collaboration and decision-making tasks, but in settings where such models cannot be called at deployment time. However, it is hard to specify the behavior of an individual robot from a team instruction, and have it continuously adapt to actions from other robots. This necessitates a framework with the representational capacity required by the logic and semantics of a task, and yet supports decentralized, real-time operation. We solve this dilemma by recognizing that a task can be represented as a deterministic finite automaton, and that recurrent neural networks (RNNs) can encode numerous automata. This allows us to distill the logic and sequential decompositions of sub-tasks obtained from a language model into an RNN, and align its internal states with the semantics of a given task. This leads to a tiny model that encapsulates the reasoning of the language model and can be implemented onboard. To interpret the internal state of the RNN for a decentralized execution, we train a graph neural network control policy conditioned on the hidden states of the RNN and the language embeddings. We present evaluations on simulated and real-world multi-robot tasks that require sequential and collaborative behavior by the team, demonstrating scalable, robust, real-time performance -- sites.google.com/view/prompting-teams.

Learning long-horizon robotic manipulation requires jointly achieving expressive behavior modeling, real-time inference, and stable execution, which remains challenging for existing generative policies. Diffusion-based approaches offer strong modeling capacity but incur high inference latency, while flow matching enables fast, near-single-step generation yet often suffers from unstable execution when operating directly in the raw action space. We propose Continuous Latent Action Flow Policy (CoLA-Flow Policy), a trajectory-level imitation learning framework that performs flow matching in a continuous latent action space. By encoding action sequences into temporally coherent latent trajectories and learning an explicit latent-space flow, CoLA-Flow Policy decouples global motion structure from low-level control noise, enabling smooth and reliable long-horizon execution. The framework further integrates geometry-aware point cloud conditioning and execution-time multimodal modulation, using visual cues as a representative modality to enhance real-world robustness. Experiments in simulation and on real robots show that CoLA-Flow Policy achieves near-single-step inference, improves trajectory smoothness by up to 93.7% and task success by up to 25 percentage points over raw action-space flow baselines, while remaining significantly faster than diffusion-based policies.

Highly dynamic motions such as running expose legged robot actuators to substantial impact torques, necessitating impact-resilient or impact-robust designs. Accordingly, actuator design must explicitly account for these impact robustness requirements. We present solutions for the impact-induced peak torque at the gearbox, where we analyze applicable contact time and parameter sensitivity. From these, we derive algebraic inequalities to determine the maximum stiffness of series elastic actuators that exhibit impact robustness. Derived from a timedomain solution of the post-impact dynamics, the proposed closed-form inequalities incorporate gearbox peak torque, reduction ratio, motor and link inertia, operational torque acting before and during impact, as well as impact impulse. We demonstrate this method by designing a humanoid actuator featuring both a novel 14 J hollow-shaft spring with 546 Nm/rad, and a conceptual axial torque-overload breakaway disk for added safety. Experimental evaluation of the impact response validates our impact-torque prediction for two different contact-time regimes, three different link inertia settings and two different stiffness settings.

In this paper, we present an online method for converting an arbitrary geometric path, represented by a sequence of states, and generated by any planner (e.g., sampling-based planners such as RRT or PRM, search-based planners such as ARA*, etc.), into a kinematically feasible, jerk-limited trajectory. The method generates a sequence of quintic/quartic splines that can be discretized at a user-specified control rate and streamed to a low-level robot controller. Our approach enables real-time adaptation to environmental changes and can be re-invoked at any instant to generate a new trajectory from the robot's current state to a desired target state or sequence of states. Under a bounded-obstacle-velocity assumption, the method provides conditional stopping-safety guarantees over a finite time interval in dynamic environments, while allowing bounded geometric deviation from the original path. Kinematic constraints, including jerk limits, are explicitly considered. We validate the approach in a comparative simulation study against a competing method, demonstrating favorable behavior w.r.t. smoothness, computational time, and real-time performance, particularly with frequent target-state changes (up to 1 [kHz]). Real-robot experiments demonstrate applicability in real-world scenarios, including scenarios with a human as an obstacle.

Traditional rigid submersibles often lack the acoustic stealth and agility required for close‐range deep‐sea exploration. While soft robotics have emerged as a promising approach for navigating complex marine environments, these systems typically face a trade‐off between pressure tolerance and practical maneuverability. We present DeepTrack, an untethered soft robotic fish driven by a compact electromagnetic propulsion system. Integrating a pressure‐independent actuation mechanism into a compliant silicone body keeps the robot's mass at just 475 g and eliminates the need for bulky pressure vessels. A two‐fin configuration decouples forward propulsion from depth regulation, enabling agile 3D maneuvering. An onboard visual servoing system ensures precise closed‐loop target tracking. Experimental validations demonstrated a maximum swimming speed of 5.78 cm/s (0.27 body length (BL)) and a yaw rate of 0.26 rad/s. During dynamic trajectory tracking, the system maintained a root mean square error (RMSE) below 0.37 BL. Open‐water field trials further validated the robot's robustness in autonomously tracking dynamic targets amid ambient currents and unstructured visual clutter. Furthermore, hyperbaric chamber tests confirm that the robot preserves its structural integrity and tracking performance at hydrostatic pressures up to 30 MPa. These results demonstrate the potential of DeepTrack as a mechatronic platform for future visually guided operations in deep sea environments.

The endurance limitations of onboard batteries and operational interruptions caused by static charging remain major barriers to scalable unmanned aerial vehicle (UAV) deployment in emerging low‐altitude economy. Dynamic wireless power transfer (DWPT) has recently emerged as a key enabling technology for persistent aerial mobility by allowing energy replenishment during flight without landing. This paper presents a comprehensive survey and introduces a unified field‐motion system framework that links DWPT mechanisms with UAV mobility states and operational requirements. Existing DWPT approaches are systematically classified along two orthogonal dimensions: physical field characteristics (near‐field inductive coupling and far‐field radiative transmission) and motion dynamics (quasi‐static hovering to fully dynamic flight). The survey reveals that near‐field ground‐to‐air configurations achieve high efficiency and robustness through Parity‐Time (PT)‐symmetric compensation and three‐dimensional coil structures, while far‐field air‐to‐air energy transfer enabled by beam‐shaped laser and microwave arrays supports long‐range energy relay under dynamic alignment uncertainty. Building upon these insights, the paper proposes a layered heterogeneous energy architecture integrating high‐efficiency local charging nodes with flexible long‐distance energy relays, forming a multi‐tier aerial energy replenishment network analogous to aerial refuelling systems. By explicitly connecting energy transfer modalities with UAV operational dynamics, infrastructure deployment, and persistent mission planning, this review establishes a system‐level perspective for energy‐aware UAV transportation networks. The presented framework provides design guidelines and research directions toward scalable, continuously operating aerial logistics, inspection, and sensing services in future low‐altitude intelligent transportation systems.

Reliable autonomous docking of AUV under ice is critical for operational safety, as maintaining vehicle integrity and achieving full‐lifecycle autonomous operations are unique challenges in ice hole recovery. This study presents a vision‐guided navigation framework specifically designed for ice hole docking, which enables accurate identification and localization of ice hole light beacons array. The framework integrates an adaptive selection algorithm that ensures robust light source detection across varying lighting conditions and operational distances, along with a dual‐quadrilateral geometry‐based recognition method capable of accurately discerning beacon positions in different visibility conditions during docking process. A monocular pose estimation algorithm provides real‐time and precise relative position and heading information, while an ice hole alignment strategy guides the AUV for safe and accurate docking. Images of six beacons arranged around the ice hole were captured by an upward‐facing camera mounted on the AUV and processed through the proposed pipeline. Extensive simulations and real‐world experiments demonstrate that the method significantly improves positioning accuracy, robustness, and overall navigation reliability, offering a practical and effective solution for underwater optical vision‐guided docking under ice.

Reliable fault diagnosis is essential for the safe operation of autonomous underwater vehicles (AUVs) in uncertain and dynamic environments. However, conventional closed‐set diagnosis methods are unable to handle previously unseen fault conditions, while existing open‐set techniques often struggle with low‐frequency, highly coupled multivariate telemetry. This paper proposes ProtoNet‐MDAG, a prototype‐based open‐set diagnosis framework for AUV fault recognition. Specifically, a dilated multi‐scale encoder is developed to extract discriminative temporal features from low‐frequency sensor streams, and a variance‐penalized prototypical learning strategy is introduced to enforce compact and well‐structured known‐class manifolds in the latent space. Based on this representation, an adaptive Mahalanobis distance gating mechanism is constructed to perform statistically calibrated open‐set rejection. Experimental results on a Haizhe AUV and a BlueROV‐class underwater robotic platform show that the proposed method achieves 97.19% and 89.12% accuracy, respectively. The results demonstrate strong unknown‐state rejection on the Haizhe AUV and competitive known‐class recognition on the BlueROV‐class platform, while also revealing that platform‐dependent distribution shift can weaken unknown‐state rejection in more heterogeneous operating conditions. Robustness tests on the Haizhe dataset further show that the framework remains stable under representative telemetry perturbations. These results demonstrate that the proposed framework provides an effective and reliable solution for open‐set fault diagnosis in underwater robotic systems.

For underwater tasks that require an efficient approach over long distances and stable adhesion for close‐range crawling, this paper proposes a propeller–leg hybrid underwater adhesion robot with integrated swimming and crawling capabilities. First, we present the overall design of a robot that integrates a bionic leg–foot mechanism and a propeller propulsion system. Second, we design a hydrogel biomimetic sucker with reversible adhesion and detachment, accompanied by an adhesion performance test. A dual‐criterion system for static stability margin and adhesion force reliability was proposed during the robot crawling stage. On the basis of computational fluid dynamics simulation results, a critical instability model was developed to quantify the effects of flow velocity, slope, and load on system stability. We have developed a crawling–swimming collaborative control strategy to address the mode‐switching issue during locomotion. This strategy employs a hierarchical control architecture guided by stability criteria, in which central pattern generator–based rhythmic control dynamically integrates with sliding‐mode propulsion compensation via switching signals. The experimental results show that the proposed robot can achieve stable attachment motion through propeller–leg coordination under special working conditions (inverted surfaces and slopes). The proposed robot provides a multimodal solution for underwater operations that require a rapid target approach in open water, followed by stable attachment and adhesive crawling on structural surfaces.

Underwater area surveillance plays a crucial role in upholding national security, safeguarding strategic interests, and mitigating potential threats. Despite its significance, a notable gap persists in the availability of comprehensive models for assessing its effectiveness. This article introduces a methodology for evaluating the operational efficiency of underwater area‐surveillance missions, with a focus on quantifying their effectiveness. By combining the availability vector, credibility matrix, and capability matrix, this approach yields evaluations of the operational efficiency of multi‐Unmanned Underwater Vehicle (multi‐UUV) systems during underwater area‐surveillance missions. Moreover, the article includes simulations of underwater area‐surveillance scenarios involving submarines, UUVs, and other targets. These simulations have pinpointed shortcomings in surveillance strategies, offering valuable insights to drive the subsequent execution, refinement, and advancement of surveillance strategies for multi‐UUV systems.