Adaptive Model-free Control with Supervising Switching Technique for Robotic Manipulator with Actuator Failure

Adaptive Model-free Control with Supervising Switching Technique for Robotic Manipulator with Actuator Failure

연구 분야: Software Development

학회: 2022 4th International Conference on Robotics and Computer Vision (ICRCV)

This paper proposes a time-delay estimator-based adaptive model-free controller with supervising switching technique for robotic manipulator with actuator failure. The proposed method based on ultra local model is composed of timedelay estimator (TDE), proportion-differential control (PDC), supervising switching technique (SST) and auto-tuning algorithm based on Nussbaum function. The TDE is adopted to estimate the lumped disturbance, while PDC is designed to stabilize the closed-loop system. For better control performance when actuator failure occurs, the SST and variable adaptive gain are introduced to suppress error induced by TDE and realize fast convergence. To validate the effectiveness and superiority of the referred strategy, compared simulations are conducted via Matlab/ Simulation. The simulation results demonstrate that the proposed controller owns high accuracy, fast convergence and less chattering in dealing with actuator failure problem.