Driver state detection for driver-automation shared control with fuzzy logic

This paper focuses on driver-automation shared lateral control by considering the variation of driver state which is interfered by multiple-risk abnormal behaviours. First, four abnormal behaviours, i.e. smoking, calling, yawning and drowsiness, under Driver Monitoring System are detected by a computer vision based method which combines face alignment algorithm with Haar–AdaBoost. A novel fuzzy logic system is then designed by analysing the risk of the four behaviours, aiming to identify the risk levels of driver states. Based on the identified results, a fuzzy inference logic is developed to design the driver-automation shared control by using a PID controller. Simulation experiments are conducted to illustrate the effectiveness of the detection method and the designed controller. The comparison between the simulation results shows that the proposed control architecture has comparatively better performance in lane keeping task.