Man-machine Collaborative Lane Keeping Control Algorithm Based On Fatigue Detection

With social progress and economic development,the number of car ownership is increasing and traffic accidents occur frequently.Lane deviation caused by driver fatigue is a common cause of traffic accidents,and lane keeping as a representative lateral assist system is an effective way to avoid lane deviation.In order to solve the above problems,the human-machine cooperative lane keeping based on fatigue detection is investigated in this paper,proposes a driver fatigue detection method based on facial feature recognition and combines it with human-machine cooperative lane keeping control,mainly accomplishing the following work.Firstly,a driver fatigue detection method based on facial feature point recognition is proposed,using a camera to capture driver facial image information.After that,Adaboost face detection algorithm and ERT feature point location algorithm were used to detect the face,extract the feature points and calculate the fatigue index.The fatigue state of the driver was determined by the multi-index fusion fatigue detection algorithm.Secondly,the vehicle kinematic model and dynamics model are established,and the trajectory tracking driver model is built on the basis of pre-scanning following and neural network theory,and the effectiveness of the driver model is verified through simulation experiments.Based on the model predictive control theory,MPC lateral controller is designed to realise the lateral position control of the vehicle and control the steering through the front wheel turning angle,and is verified through simulation experiments.Based on fuzzy control,the human-machine cooperative weight allocation strategy is designed to dynamically adjust the human-machine control weights according to the driver fatigue index and lane deviation from the centre line,and to change the intervention level of the automatic controller to reduce the lateral deviation of the vehicle,so as not to interfere excessively with the driver’s operation on the premise of ensuring the control accuracy.Finally,hardware-in-the-loop experiment were used to verify the control strategy.The results show that the human-machine weight controller can dynamically adjust the driver’s fatigue state and lane departure condition,reduce the human-machine conflict,avoid lane departure and realize the human-machine cooperative lane keeping on the basis of retaining the driver’s certain operation.