论文部分内容阅读
A novel adaptive dual network design consisting of a rough adjustment network (RAN) and a fine adjustment network (FAN) is proposed to eliminate the unknown time-variant uncertainties of servo system. To accomplish this objective, a RAN is proposed based on the combination of sliding mode control, function approximation, and error compensation technique. Then, an FAN is proposed to compensate the tracking error. In our current design, the FAN includes a critic network based on a neural network model and a prediction network based on an online curve fitting scheme. Theoretical analysis followed by detailed design strategies are presented in this work. Simulation results and comparative study of this method with those of existing approaches demonstrate the effectiveness of the proposed adaptive dual network design for position tracking.
A novel adaptive dual network design consisting of a rough adjustment network (RAN) and a fine adjustment network (FAN) is proposed to eliminate the unknown time-variant uncertainties of servo system. To accomplish this objective, a RAN is proposed based on the combination of sliding mode control, function approximation, and error compensation technique. Then, an FAN is proposed to compensate the tracking error. In our current design, the FAN includes a critic network based on a neural network model and a prediction network based on an online curve fitting scheme. Theoretical analysis followed by detailed design strategies are presented in this work. Simulation results and comparative study of this method with those of existingathered demonstrates the effectiveness of the proposed adaptive dual network design for position tracking.