A5017639281- Adaptive Dynamic Programming Control
- Prosthetics and Rehabilitation Robotics
- Adaptive Control of Nonlinear Systems
- Stroke Rehabilitation and Recovery
- Reinforcement Learning in Robotics
- Distributed Control Multi-Agent Systems
- Neural Networks Stability and Synchronization
- Muscle activation and electromyography studies
- Frequency Control in Power Systems
- Mechanical Circulatory Support Devices
- Stability and Control of Uncertain Systems
- Advanced Control Systems Optimization
- Fault Detection and Control Systems
- Viral Infections and Vectors
- EEG and Brain-Computer Interfaces
- Drilling and Well Engineering
- Spinal Cord Injury Research
- Fuzzy Logic and Control Systems
- Tunneling and Rock Mechanics
- Iterative Learning Control Systems
- Smart Grid Security and Resilience
- Balance, Gait, and Falls Prevention
- Fuzzy Systems and Optimization
- Robotic Locomotion and Control
- Network Security and Intrusion Detection
University of Electronic Science and Technology of China
2018-2025
In this article, a novel reinforcement learning (RL) method is developed to solve the optimal tracking control problem of unknown nonlinear multiagent systems (MASs). Different from representative RL-based algorithms, an internal reinforce Q-learning (IrQ-L) proposed, in which reward (IRR) function introduced for each agent improve its capability receiving more long-term information local environment. IrQL designs, Q-function defined on basis IRR and iterative algorithm learn optimally...
In this paper, an event-triggered optimal tracking control of discrete-time multi-agent systems is addressed by using reinforcement learning. contrast to traditional learning-based methods for coordination and with a time-triggered mechanism, mechanism proposed update the controller only when designed events are triggered, which reduces computational burden transmission load. The stability analysis closed-loop described. Further, implement scheme, actor-critic neural network learning...
This paper reviews recent progress in model identification-based learning and optimal control its applications to multi-agent systems (MASs). First, a class of learning-based method, namely adaptive dynamic programming (ADP), is introduced, the existing results using ADP methods solve problems are reviewed. Then, this investigates various kinds identification analyzes feasibility combining method with unknown systems. In addition, expounds current fields single-agent (SASs) MASs. Finally,...
This article investigates an output antisynchronization problem of multiagent systems by using input-output data-based reinforcement learning approach. Till now, most the existing results on problems required full-state information and exact system dynamics in controller design, which is always invalid practical scenarios. To address this issue, a new representation constructed just available input/output data from system. Then, novel value iteration algorithm proposed to compute optimal...
ABSTRACT This article mainly presents a fresh systematic framework to tackle the resilient fault‐tolerant sampled‐data control (SDC) synthesis problem for networked interval type‐2 fuzzy systems (IT‐2FSs) suffering with semi‐Markovian‐type jump actuator failures (SMJAFs) and mismatched membership functions (MMFs), which portrays more features than some prior developments. The principally target of addressed under this investigation is precisely architect faulty mode‐dependent controller such...
Lower limb exoskeleton (LLE) has received considerable interests in strength augmentation, rehabilitation and walking assistance scenarios. For assistance, the LLE is expected to have capability of controlling affected leg track unaffected leg's motion naturally. An important issue this scenario that system needs deal with unpredictable disturbance from patient, which requires controller ability adapt different wearers. This paper proposes a novel Data-Driven Reinforcement Learning (DDRL)...
In this brief, a tracking control problem for robotic systems with unknown uncertainties is addressed by using an event-triggered adaptive dynamic programming (ADP) method. First, the of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$n$ </tex-math></inline-formula> -degree freedom (DOF) system transformed to optimal auxiliary such that robust design original feasible based on ADP framework. To reduce...
More recently, lower limb exoskeletons (LLE) have gained considerable interests in strength augmentation, rehabilitation, and walking assistance scenarios. For assistance, the LLE is expected to control affected leg track unaffected leg's motion naturally. A critical issue this scenario that exoskeleton system needs deal with unpredictable disturbance from patient, controller has ability adapt different wearers. To end, a novel data-driven optimal (DDOC) strategy proposed hemiplegic patients...
A challenging task for the biological neural signal-based human-exoskeleton interface is to achieve accurate lower limb movement prediction of patients with hemiplegia in rehabilitation training scenarios. The based on single-modal signals such as electroencephalogram (EEG) currently not mature predicting movements, due its unreliability. multimodal a very novel solution this problem. This kind normally combines EEG signal surface electromyography (sEMG) signal. However, their use still...
The human–robot interface (HRI) based on biological signals can realize the natural interaction between human and robot. It has been widely used in exoskeleton robots recently to help predict wearer's movement. Surface electromyography (sEMG)-based HRI mature applications exoskeleton. However, sEMG of paraplegic patients' lower limbs are weak, which means that most limb cannot be applied Few studies have explored possibility using upper In addition, HRIs do not consider contribution synergy...