PID 제어기와 신경망 보상기를 사용한 로봇매니퓰레이터 제어

Control of Robot Manipulator using PID Controller & Compensator of Neural Network

This paper introduces an adaptive compensator using a neural network and PID controller without the plant model. The reference input of a PID controller is converted to a newly learned input by a neural network compensator. The new reference input considers the situation between existing disturbance and time-varying plant parameters. The learned input assists the plant output to trace the reference input properly. The multilayer neural network is employed for the reference compensator. The back error propagation algorithm is used to minimize the errors between the reference and plant outputs. The simulation results show that the proposed method is more robust than a PID controller with fixed gain.