Abstract: An adaptive directional control driver model is established to simulate the driver’s learning behavior about vehicle dynamic characteristic. Drivers are divided into eligible group and professional group according to their directional control skill. During the learning process, novice drivers will have a basic understanding of vehicle dynamics after low speed slow steering training, so they can proficiently control the vehicle at low speed range. A direction control model based on genetic algorithm offline tuning is built to simulate this learning process from a novice driver to an eligible driver in light of the characteristics of driver’s control behavior at low speed range. The dynamic response of vehicle shows a strong nonlinear property when the lateral acceleration or the speed is high, eligible drivers need to gather the driving experience at high speed range to control the vehicle skillfully, finally becoming an expert driver. So, for strong nonlinear characteristic of the vehicle at high speed, neural network online tuning is added to the original model to simulate learning process from an eligible driver to a professional driver under high-speed condition. The simulation results of model are well consistent with the behavior of real driver.

Key words: Direction adaptive control, Driver model, Genetic algorithm, High-speed nonlinear, Neural network