보상용 신경회로망을 사용한 이동로봇의 속도제어

Velocity Control of Mobile Robot using Neural Network for Compensation

It is very difficult problem to implement stable trajectory tracking algorithms because mobile robots have kinematically non-holonomic constraints. For solving this problem, a velocity controller is presented in this paper. The velocity errors caused by system uncertainties or internal and external disturbances exist. A neural network is used for compensating the velocity errors. Input variables of this neural network compensator are defined by differences between the velocities of the posture controller and the real velocities of the mobile robot. The controller containing the compensator is more robust against the system uncertainties and disturbances as compared with the case which just uses a PID controller.