Impedance Control Based on Self-Learning Parameters of Neural Network
Aiming at the problem that the robot's environmental stiffness and position are unknown in the grinding process and the traditional impedance control is difficult to maintain the grinding quality effective-ly,an impedance control based on the self-learning of neural network with adjustment parameters is pro-posed.Since the selection of the regulation parameters in the damping parameter compensation method de-signed based on Lyapunov stability theory directly affects the control performance of the system,a neural network is constructed for its parameter adaptive adjustment based on the mathematical description of damping compensation,and different excitation functions are designed to reflect the characteristics of damp-ing changes under the influence of multiple factors.Through the built neural network online learning,the optimization of parameters is realized to adapt to the environmental changes of the grinding process.Simu-lation experiments in different environments,such as inclined,flat and curved surfaces,considering their sudden changes in stiffness,dynamic changes in stiffness and dynamic changes in expected force,show that the proposed control method has smaller overshoot and steady-state error compared with the traditional con-trol method,and can adapt to the situation where the environmental parameters are unknown,significantly improving the grinding quality and efficiency.
NETL
NSTL
万方数据
维普
