A driver assistance framework based on driver drowsiness detection

This paper proposes the development of a driver assistance framework that allows switching between manual and autonomous driving on a simulated testbed. This testbed is a driving simulator which can create realistic environments and driving experience while enabling both autonomous and manual driving capabilities. The switching between manual and autonomous driving is triggered by the driver drowsiness analysis which is performed with two inputs: driver's facial data (from a camera mounted in front of the driver) and steering wheel data (from a car controller system). In this paper, the driver status is always monitored. When he is not drowsy, he can manually drive the simulated car. Otherwise, the simulator lets the car switch to autonomous driving to run in a pre-defined trajectory. Experiments were performed on the simulated testbed to demonstrate the effectiveness of the proposed framework.