Vehicle drivers play an important role in transportation safety and vehicle design. Understanding the driver’s behavior, especially the impaired driver’s behavior, is crucial to improve the vehicle safety. The proposed impaired driver model is based on the optimal preview control and the linear quadratic regulator. Two important parameters that could be counted for in a mathematical model of the driver are the reaction time and the preview time. For the impaired driver model, the value of reaction time is increased while the value of preview time is decreased. The simulation results for the model of the impaired driver and the vehicle produce a larger lateral deviation than the one of a normal driver, as revealed in the experiments conducted by the previous studies. The investigation on vehicle parameters reveals that the changes of parameters may improve the overall performance of the impaired driver-vehicle system. The controller for autonomous vehicles developed from the studies of the driver model may eliminate the negative effect of impaired drivers. The preview capability of driver is introduced to the design of the controller by using the preview control theory. The preview information of the path in terms of the lateral position and the velocity profile enhances the performance of the autonomous vehicle. The neural network is presented as a feasible alternative approach to implement the future path in design of autonomous vehicle controller. The neural network weights the path data and provides the adjustment as the preparation to the vehicle.
展开▼