Identification of abnormal driving state based on driver's model

An abnormal driving state recognition method is proposed in this paper. The basic idea is to regard the driving behavior as a feed back control system; the input signal is vehicle lateral position; the output signal is steering wheel angle. The control system, i.e., the driving behavior, is modeled with normal driving data. And then the model is used to fit the driving data of unknown driving state. The fitting error was evaluated and monitored with statistical process control (SPC) approach, i.e., the control charts tools. If the fitting error is evidently different from that of normal driving state, then abnormal driving state is recognized. Experiment results verified the effectiveness of the proposed method.