Research On Vehicle Following Control Algorithm Based On Road Adhesion Level And Driving Style

With the improvement of economic level,automobile has become an important means of travel,followed by the frequent occurrence of traffic accidents.In view of this problem,vehicle assisted driving technology has been developed rapidly.Vehicle assisted driving technology can effectively alleviate driving fatigue and reduce the occurrence of traffic accidents.Vehicle following system is an important product in the practical application of intelligent assistant driving,which can reduce the driver's driving fatigue in the process of following the vehicle,reduce the occurrence of traffic accidents such as rear end collision,and become one of the key technologies to improve the intelligent automobile.In this paper,combined with an engineering research and development project,by summarizing and analyzing the research status of vehicle following system at home and abroad,the estimation of road adhesion level and the recognition of driving style are studied,and the vehicle following control algorithm based on model predictive control is designed.The main research contents include the following aspects:Firstly,the paper studies the recognition algorithm of road adhesion grade considering road slope.Based on the longitudinal dynamic model of vehicle,the recursive least square method with forgetting factor is used to build the algorithm model of road slope estimation with vehicle speed,acceleration and engine torque as input variables;the calculation formulas of wheel ground load and driving wheel longitudinal force are derived,and the algorithm model of adhesion coefficient estimation is built based on the recursive least square algorithm;the recursive Kalman method is used to build the algorithm model of adhesion coefficient estimation Filtering algorithm,taking the difference between driving wheel and driven speed as parameters,constructs the algorithm model of slip rate estimation of driving wheel;revises the mathematical model of slip rate-using the coefficient of adhesion,and constructs the algorithm model of on-line estimation of road adhesion coefficient based on Bayesian principle;builds MATLAB/ Simulink and CarSim combined simulation platform,the road slope estimation algorithm and road adhesion coefficient estimation algorithm are simulated and verified.Secondly,based on the K-means clustering analysis method,the driving style recognition algorithm is studied.The real vehicle data acquisition system is built,and the real vehicle test under the urban expressway traffic environment is designed and completed,and the effective driving test data is obtained;the driving style type and identification cycle length are defined,and the driving style identification index is determined according to the dynamic and kinematic parameters during the driving process;the original index is reduced by using the principal component analysis method Based on the k-means algorithm,the driving style online recognition model is constructed.Thirdly,a multi-objective optimal vehicle following control algorithm is designed based on the model prediction method.Based on the actual car following data,a safe car distance model considering the road adhesion level and driving style is designed,and the kinematic model of the vehicle following control system is constructed,and the state space equation is determined;the constraint conditions of the model predictive control are determined according to the selected design indexes of safety,following and comfort;the predictive equation and performance of the model predictive control are derived The model predictive control(MPC)is transformed into a linear programming problem with parameters such as index function.The design of vehicle following control algorithm based on MPC is completed.Fourthly,the simulation analysis of typical vehicle following condition and the real vehicle test verification of the algorithm of road slope and road adhesion grade estimation are carried out.A joint simulation platform based on CarSim and MATLAB / Simulink is built to simulate and analyze typical working conditions;a real vehicle test platform is built to verify the road slope estimation algorithm and the road adhesion level estimation algorithm.Simulation analysis and real vehicle test results verify the effectiveness and practicability of the algorithm.