1993 Volume 55 Issue 5 Pages 83-92
The object of this study is to control agricultural vehicle utilizing neural network system (NNS) through nonlinear controller. Kinematics of the vehicle was described as the emulator, that was nonlinear dynamic model, by NNS, and its accuracy was investigated using the training data. The generalization of the emulator for the vehicle was also confirmed by the movement patterns which were not the training data. It was made clear that the output of NNS were in good agreement with the experimental data, and were correcter than the output of the physical model neglecting slip angle. Furthermore, the simpler movements such as lane change and meandering motion were carried out to evaluate the ability of the emulator.
