Research On The Personalized Adaptive Cruise Control System With Driver Style Identification

Adaptive Cruise Control(ACC)is an advanced driver assistance system(ADAS)that enables steady cruising,safe controlling,and car following,which could effectively alleviate driving fatigue,reduce traffic accidents and improve traffic flow.However,ACC systems nowadays are generally designed in the uniform mode,lacking the consideration of the drivers’ actual needs.As a matter of fact,driver behavior is significantly different between different types of drivers,the requirements of car-following are quite different between different types of drivers.Aggressive drivers tend to follow target vehicles more closely and quickly,however,moderate drivers tend to keep large inter-vehicle distance and relatively gentle acceleration and braking operation.Therefore,clarify the driver behavior characteristics,embed it in ACC system is the important research direction and the key generic technology of intelligent vehicles in the future.Supported by the National Natural Science Foundation of China,this paper designs a personalized adaptive cruise control system with driver style identification.Firstly,two driver behavior data acquisition systems are established,including the driver in the loop simulator and a real vehicle driving data acquisition platform.Based on which,the driver behavior data acquisition tests are respectively carried out under the typical working conditions.The driving data are analyzed and filtered through factor analysis,cluster analysis,and then,a three levels back propagation neural network model is established to identification the driver behavior characteristics offline and online.Then,the personalized adaptive cruise control system is degisned with the identification results of driver behavior.Finally,simulations and real vehicle tests are carried out to verify and analyze the driver behavior identification stragety and personalized adaptive cruise control system.This paper carried out the following tasks,(1)Driver behavior data acquisitionThe driver behavior data acquisition systems include a driving simulator and a real vehicle driving data acquisition platform.The driving simulator is used to carry out the high-speed driving behavior data acquisition tests.The real vehicle driving data acquisition platform is used to carry out the driver behavior data acquisition tests under the typical ACC working conditions.(2)Research on the analysis and identification of driver behavior characteristicsThe factor analysis is used to simply the data dimensions and find feature parameters,and the driver style classification is processed by K-means algorithm.Then,a three levels BP neural network model is established to realize the driver behavior characteristics identification online.(3)Research on the personalized adaptive cruise control strategyThe driver behavior data of typical ACC working conditions are analized.Time headway,mode switching logic and the personalized desiered acceleration are used to represent the difference among different drivers.The personalized parameters are regulated online based on the driver style identification results.(4)The verification of the identification and the control strategyThe real-time driver behavior characteristics identification strategy are carried out by the driving simulator and the real vehicle tests.And personalized adaptive cruise control system is verified by the co-simulation platform of MATLAB/Simulink and Pano Sim.Finally,the driving data outputted by personalized adaptive cruise control system and the real driving data of different drivers are compared,The results show that the proposed real-time driver behavior identification stragety has a high accuraty and precision and Personalized adaptive cruise control system could effectively adapt to the current driver’s driving style,and satisfy the safety driving,car-following,and ride comfort performance.