Acquisiton of Car Cornering Operation by Means of Neural Networks and Genetic Algorithms.

Nowadays, R&D of intelligent transport system(ITS)is actively performed, and the aid system of driving operation is becoming important. The aid system of driving operation has to be designed with consideration of characteristics of car drivers. The system designed without such consideration gives strange feeling to drivers and conflicts with drivers. In this paper, a nonlinear steering controller is designed so that it has a similar structure to that of car drivers. At usual situation, the characteristics of car motion are regarded to be almost linear, and the drivers' operation and the control structure are simple enough. On the other hand, at a quick motion situation, the characteristics of car motion become nonlinear, and the drivers' operation and the control structure should be treated to be nonlinear. Therefore, a neural network(NN)steering controller that consists of 3 NNs is applied to such a problem. This structure has a similar function and structure to that of car drivers. These NNs are learned by using the genetic algorithm with numerical simulation for the case of passing through some simple curved roads. The effectiveness and the feasibility of this controller are verified by numerical simulation.