Chen Wang

This paper proposes a neural‐network‐based adaptive fixed‐time fault‐tolerant control (AFTFTC) scheme for robotic manipulators with unknown dynamics and actuator faults. A radial basis function neural network (RBFNN) is first employed to approximate the unknown nonlinear dynamics, an adaptive law is designed to estimate an unknown scalar parameter associated with the norm of the ideal neural‐network weights. To enhance the fault‐tolerance capability, an additional adaptive law is developed to compensate for unknown actuator faults. By integrating these adaptive mechanisms into the backstepping control framework, the proposed AFTFTC controller is constructed to handle both normal and faulty operating conditions. Theoretical analysis shows that both the joint tracking errors and the adaptive parameter estimation errors converge to a small neighborhood of zero within a fixed time, independent of the initial conditions. Comparative experiments on a robotic manipulator are carried out to verify the improved tracking accuracy, robustness, and fault‐tolerant performance of the proposed method.

In this study, an adaptive dynamic programming (ADP)-based optimal backstepping control (OBC) method is proposed for robotic manipulators with unknown dynamics. The original second-order dynamic model is first decomposed into two subsystems, namely, the position-loop subsystem and the velocity-loop subsystem. Approximate optimal controllers are then designed for each subsystem using the ADP framework, in which the update laws of the actor and critic neural network weights are derived based on a positive definite function. In addition, an identifier neural network is developed to compensate for the unknown nonlinearities arising in the control design process. On this basis, an ADP-based OBC scheme is established for manipulator motion control. Rigorous theoretical analysis proves that the closed-loop system achieves not only uniformly ultimately bounded (UUB) stability but also near-optimal control performance. Finally, the proposed method is experimentally validated through a static target pick-and-place task, and the results confirm its effectiveness and superior control performance.

This study investigates reinforcement-learning-based fixed-time optimal formation control for multiple nonholonomic mobile robots with prescribed performance constraints. First, the constrained formation error dynamics is established using a leader-follower approach. Meanwhile, a barrier function is employed to transform the constrained formation error dynamics into an unconstrained form. Then, an adaptive control technique and a critic-only reinforcement learning strategy are utilized to design a fixed-time optimal control law for the unconstrained error dynamics. Rigorous theoretical derivations demonstrate that the proposed control law guarantees that the constrained formation error converges to near zero within a fixed time, optimizing the performance index while satisfying the prescribed performance requirement. Finally, the feasibility of the proposed method is verified through simulations and experiments.

In this study, the adaptive dynamic programming (ADP)-based fixed-time optimal trajectory tracking control is investigated for wheeled mobile robots. An ADP-based fixed-time optimal tracking controller is developed based on the critic-only neural network ADP technique, which guarantees the robot track the desired trajectory in fixed time. Firstly, to address the solution difficulty of the Hamilton-Jacobi-Bellman (HJB) equation, a critic neural network is used to estimate the cost function. Meanwhile, a weight update law is designed by using the adaptive control technique, which not only removes the persistent or finite excitation condition, but also enables the fixed-time convergence of the weight estimation error. By using the proposed controller, all error variables can converge to a neighborhood of zero in fixed time. Finally, both simulations and physical experiments indicate that the proposed ADP-based fixed-time optimal controller has a faster convergence rate compared to the two comparison controllers.

A multilevel control strategy is proposed for a lower limb exoskeleton to realize different human training modes. In the high-level control layer, the human training mode is decided by the operator’s motion intention to generate the reference gait trajectory. Meanwhile, both the joint estimation torque by the long short-term memory (LSTM) network and the human–exoskeleton interactive torques are used to evaluate the wearable comfort performance of the operator. In the middle-level control layer, a variable admittance controller is designed to plan three training modes of human–exoskeleton cooperative motion: passive, active, and passive-to-active mode (PAM). In the low-level control loop, a fixed-time convergent controller with radial basis function neural network (RBFNN) estimation law and input deadzone compensation is presented to guarantee the exoskeleton joint position tracks the desired trajectory of the admittance loop output. To avoid the Zeno phenomenon of the designed controller, an event-triggered mechanism (ETM) is used to determine the execution time for sampling and transmitting signals. Finally, the effectiveness of the proposed control strategy is verified by both simulation and experimental results.