Chapter four - Driver behaviour recognition based on head movement

Abstract

The take-over process from automatic to manual control plays a critical role in the level 3 automated driving defined by Society of Automotive Engineering, and the driver’s state prior to the transfer of control process has a great influence on the take-over quality. Estimation of driver’s attention level in various non-driving activities (NDAs) is therefore of great importance to the design of the take-over control strategy. Considering playing a tablet as a specific NDA, this chapter introduces a novel approach to characterise the driver’s attention level using a sliding-window correlation analysis on the 3D movements of the driver’s head and tablet. The correspondence between the attention level and the estimated correlation strength with time delay is then established by applying the support vector machine method. The experiment results on both stationary and non-stationary data demonstrate that the correlation of movements has the potential to act as a proxy measurement of the driver attention level if this relative movement exists.