Adaptive Speed Control of a Robot Vehicle by Neural Networks
Masami IWATSUKI, Minoru KODAIRA, Takao OHUCHI
This paper proposes an adaptive speed control method for a robot vehicle of which dynamics greatly varies by the steering angle. The proposed system consists of the PID controller and two neural networks, which tune not only a set of PID gain parameters but also a feed-forward compensation. These PID gain tuner and feed-forward compensator generate the adequate PID gains and offset according to steering angles and target speeds.
A computer simulation of vehicle motion is carried out to. show the effectiveness of a proposed control method.