As a first step towards a dynamic control-oriented driver model that can describe various levels of driving skill in the driver’s primary vehicle control functions, this paper attempts to recognize the driving skill manifested in the low-level control characteristics. Vehicle tests on a DLC course were conducted with twenty subjects of different skill levels. A variance-based analysis was applied to the steering angle to generate both the frequency responses and the time sequences for frequency-domain and time-domain analysis. Two entropy measures are developed to evaluate the effectiveness of drivers’ steering control, and preview errors are used to provide an absolute measure for drivers’ preview performance. Simple rules using these three measures are developed, and data analysis verifies that the above three measures can effectively distinguish the driving skill levels of the twenty subjects participated in the DLC vehicle test.

This content is only available via PDF.
You do not currently have access to this content.