A5103176848- Particle physics theoretical and experimental studies
- High-Energy Particle Collisions Research
- Quantum Chromodynamics and Particle Interactions
- Robot Manipulation and Learning
- Robotic Mechanisms and Dynamics
- Robotic Locomotion and Control
- Particle Detector Development and Performance
- Dark Matter and Cosmic Phenomena
- Prosthetics and Rehabilitation Robotics
- Soft Robotics and Applications
- Robotic Path Planning Algorithms
- Cosmology and Gravitation Theories
- Muscle activation and electromyography studies
- Neutrino Physics Research
- Tactile and Sensory Interactions
- Manufacturing Process and Optimization
- Genomics and Phylogenetic Studies
- Robotics and Sensor-Based Localization
- Control and Dynamics of Mobile Robots
- Reinforcement Learning in Robotics
- Teleoperation and Haptic Systems
- Computational Geometry and Mesh Generation
- Black Holes and Theoretical Physics
- Computational Physics and Python Applications
- Advanced Numerical Analysis Techniques
Tencent (China)
2018-2024
Shanghai Jiao Tong University
2005-2023
Ruijin Hospital
2023
Shanghai Institute of Hematology
2023
Meta (United States)
2023
Purdue University West Lafayette
2010-2022
Southwest Jiaotong University
2020-2022
Defence Electronics Research Laboratory
2022
National University of Singapore
2008-2021
Lanzhou University of Technology
2020-2021
This letter studies the adaptive optimal control problem for a wheel-legged robot in absence of an accurate dynamic model. A crucial strategy is to exploit recent advances reinforcement learning (RL) and programming (ADP) derive learning-based solution control. It shown that suboptimal controllers can be learned directly from input-state data collected along trajectories robot. Rigorous proofs convergence novel data-driven value iteration (VI) algorithm stability closed-loop system are...
This paper presents a balance control technique for novel wheel-legged robot. We first derive dynamic model of the robot and then apply linear feedback controller based on output regulation quadratic regulator (LQR) methods to maintain standing ground without moving backward forward mightily. To take into account nonlinearities obtain large domain stability, nonlinear interconnection damping assignment - passivity-based (IDA-PBC) method is exploited in more general scenarios. Physical...
Human-like features, like toe-off, heel-strike, and disturbance rejection, can enhance the performance of bipedal robots. However, required control strategies for these motions influence each other, few studies have considered them simultaneously. Humans walk stably with toe-off heel-strike even after experiencing disturbances. Thus, we study human strategies, then, apply to a robot. This paper proposes bioinspired method realize stable walking robot First, analyze obtain some strategies....
Redundant manipulators play a critical role in industry and academia, which can be controlled from the kinematic or dynamic perspective. The motion-force control of redundant is core problem robot control, especially for task requiring keeping contact with objectives, such as cutting, polishing, deburring, etc. However, when manipulator's model structure unknown, it challenging to take manipulators. This article proposes data-driven-based scheme, solves scheme effect estimate information,...
Centroidal dynamics, which describes the overall linear and angular motion of a robot, is often used in locomotion generation control legged robots. However, equation centroidal dynamics contains nonlinear terms mainly caused by robot's needs to be linearized for deriving model-predictive controller. This paper proposes new linearization dynamics. By expressing with exponential coordinates, more are identified retained than existing methods reduce loss from model linearization. As...
Friction uncertainty and contact position may have a disastrous effect on the closure properties of grasps. This paper reflects our approach to handling these uncertainties in force-closure analysis. The former is measured by possible reduction rate κ friction coefficients, while radius ρ regions used quantify latter uncertainty. actual point deviate from desired but not farther than S , supremum without loss force-closure, indicates grasp tolerance For investigating above systematically, we...
This paper presents an efficient algorithm to compute the minimum of largest wrenches that a grasp can resist over all wrench directions with limited contact forces, which equals distance from origin space boundary set. value has been used as important quality measure in optimal planning for two decades, but there no way it until now. The proposed starts polytope containing set and iteratively grows such quickly converges aforementioned value. superior efficiency accuracy this previous...
In this article, a simultaneous learning and control scheme built on the joint velocity level with physical constraints decision variable its derivative, i.e., angle, velocity, acceleration constraints, is proposed for redundant manipulator control. The works when structure parameters involved in forward kinematics are unknown or implicit. parts incorporated simultaneously scheme, which finally formulated as quadratic programming problem solved by devised recurrent neural network (RNN)....
Living objects are difficult to grasp since they can actively elude capture by adopting adversarial behaviors that extremely hard model or predict. In this case, an inappropriately strong contact force may hurt the struggling living and a grasping algorithm minimize whenever possible is required. To solve challenging task, in article, we present reinforcement-learning (RL)-based with two stages: pregrasp stage in-hand stage. stage, robot focuses on object's behavior approaches it reliable...
To enrich legged robots with fast energy-efficient mobility on even terrain, wheeled-legged have emerged as a valued robot form in robotics research. This paper describes the complete development of new quadruped named Max, ranging from its mechanical design over system architecture to core algorithms implemented for it realize various motion behaviors. Instead attaching wheels distal ends legs existing designs, this has installed knees special switching mechanism convert leg between and...
Grasping and manipulating various kinds of objects cooperatively is the core skill a dual-arm robot when deployed as an autonomous agent in human-centered environment. This requires fully exploiting robot's versatility dexterity. In this work, we propose general framework for manipulators that contains two correlative modules. The learning-based dexterity-reachability-aware perception module deals with vision-based bimanual grasping. It employs end-to-end evaluation network probabilistic...
Uneven terrain walking is one of the key challenges in bipedal walking. In this paper, we propose a motion pattern generator for slope 3D dynamics using preview control zero moment point (ZMP). method, future ZMP locations are selected with respect to known gradient. The trajectory center mass (CoM) robot generated by controller maintain at desired location. Two models walking, namely upslope and downslope, investigated. Continuous on slopes different gradients also studied enable robots...
Electrical impedance tomography (EIT) is a technique for reconstructing the conductivity distribution by injecting currents at boundary of subject and measuring resulting changes in voltage. Image reconstruction EIT nonlinear ill-posed inverse problem. The Tikhonov method with L2 regularization always used to solve However, smoothes sharp or discontinue areas reconstruction. using L1 allows addressing this difficulty. In paper, sum absolute values substituted squares form regularization,...
The mechanisms of locomotion in mammals have been extensively studied and inspire the related researches on designing control architectures for legged robots. Reinforcement learning (RL) is a promising approach allowing robots to automatically learn policies. However, careful reward-function adjustments are often required via trial-and-error until achieving desired behavior, as RL policy behaviors sensitive rewards. In this paper, we draw inspiration from rhythmic animals propose new...
Pose estimation problems are fundamental in robotics. Most of these challenging due to the nonconvex nature. This also sets up an obstacle for uncertainty description that is essential pose integration and quality control. In this article, we show a large class related can be categorized as quadratic (QPEPs) propose general quaternion-based mathematical model unify problems. To solve QPEPs, Gröbner-basis method investigated derive their globally optimal robust solutions. Furthermore, develop...
This paper presents an algorithm to compute the minimum distance from a point convex cone in <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> -dimensional space. The is represented as set of all nonnegative combinations given set. generates sequence simplicial cones cone, such that their distances single converge desired distance. In many cases, generated finite, and therefore, has finite-convergence property. Recursive formulas are derived...
This paper presents a tracking controller considering contact force constraints for floating-base humanoid robots. The goal of the is to compute joint torques such that robot can imitate given reference motions obtained from, example, human motion capture data. technical challenge depends not only on but also forces from environment, which depend and are subject friction forces. Therefore, computing associated typically results in nonlinear optimization problem with complex constraints. We...
Safety and efficiency are two crucial factors for human-robot collaboration. It is challenging to ensure human safety while not sacrificing the task efficiency. In this letter, we present a reinforcement learning (RL) based method with hazard estimator balance these factors. Our has phases. training phase, an RL control policy trained; in testing dynamically select guiding goal along given path between avoidance execution. The proposed compared among three previous methods: another method,...
The wheel-bipedal robot has the advantages of both wheeled robots and legged robots, but as a cost, it is more challenging to perform flexible movements in various surroundings while keeping balanced. inaccurate dynamics makes balance problem even intractable. To solve this problem, Ollie used testbed. whole-body control (WBC) framework adopted enhance dexterity with multiple degrees freedom task space. Moreover, learning-based adaptive technique applied assist WBC such that controller can...
In combination with a robotic arm, mobile platform can fulfill complicated functions but may encounter singularity problems. To solve them, this work investigates the kinematic control of arm (MRA) that is based on redundant mounted robot. Specifically, quadratic program-based biobjective (QPBO) scheme proposed, which integrates manipulability optimization and minimum velocity infinite-norm. Accordingly, modified neural dynamics (MND) founded its stability substantiated by Lyapunov analyses...
Physical human-robot interaction (pHRI) is widely needed in many fields, such as industrial manipulation, home services, and medical rehabilitation, puts higher demands on the safety of robots. Due to uncertainty working environment, pHRI may receive unexpected impact interference, which affects smoothness task execution. The commonly used linear admittance control (L-AC) can cope well with high-frequency small-amplitude noise, but for medium-frequency high-intensity impact, effect not good....