NEURAL NETWORK CONTROL OF A VEHICLE 4WS NONLINEAR MODEL

The nonlinear characteristics of tire were taken into account in this paper, and Magic formula was adopted in order to establish the nonlinear vehicle model because the difference between the side slip force of the linear model and nonlinear model is big while the side slip angle at 12°～16°. It is proved that artificial neural network can simulate any nonlinear mathematic model with any precision and it is fit for identifying the Magic formula. This paper adopted a forward BP artificial neural network with 4 layers. The neuron quantities of input layer and output layer are 4 and 3 respectively and the first and second hidden layers had 10 and 8 neurons. The effect of identification is very good. The discrete control system of 4 wheel steering (4WS) vehicle was designed on the basis of ANN identification. This system is a negative feedback control system and PID controller was used into this system to control the front wheel angle in order that the system can reach the aim of the control system which is to minimize the side slip angle to zero. Vehicle velocity is the key influencing factor of vehicle stability. The side slip angle and the yaw rate at 80?km/h～120?km/h were calculated in the paper. The side slip angles at 80?km/h and 100?km/h equaled zero approximately and much less than the one at 120?km/h. The yaw rates at 80?km/h and 100?km/h were much less than the one at 120?km/h. The simulation results show that vehicle stability declined with vehicle velocity increasing and the control system can reach the aim of 4WS vehicle and the nonlinear model can improve manipulation and stability of vehicle effectively.