Research on inverse dynamics of FSAE racing car based on recurrent neural network

In order to research on inverse dynamics of FSAE racing car. The multi-body dynamic model of a certain FSAE racing car with 47 DOF was built, and its accuracy was verified by experiment data. Taken double lane change condition for example, the nonlinear mapping relation between lateral acceleration, velocity and steering angle was built by recurrent Elman neural network. The identification result shows, the method to study on automobile inverse handling dynamics by Elman neural network is feasible which has a rapid learning speed and high accuracy. The method can accurately identify the handling input of racing car when it has ideal performance.