Decision And Control Of High Speed Emergency Lane Change Obstacle Avoidance Of Intelligent Vehicle Considering Road Adhesion Conditions

With the continuous growth of global car ownership,vehicle active obstacle avoidance technology has attracted more and more researchers’ attention.At present,the development of Automatic Emergency Braking(AEB system is relatively mature.It can actively or assist the driver to control the vehicle brake in case of emergency to avoid the collision between the vehicle and obstacles.However,Automatic Emergency Steering(AES has higher obstacle avoidance efficiency than AEB,which requires a longer longitudinal braking distance.Therefore,from the perspective of further improving the obstacle avoidance ability of intelligent vehicles under emergency conditions,this paper focuses on emergency steering and lane change obstacle avoidance.The specific work is as follows:(1 A road adhesion coefficient estimator based on BP neural network is established.Firstly,the magic formula tire model is established;In order to avoid the problems of high difficulty and heavy workload in sample collection of real vehicle under all working conditions,a sample set for training BP neural network is constructed based on magic formula tire model;Finally,a BP neural network pavement adhesion coefficient estimator with two hidden layers is established.The test results show that the estimation accuracy of the pavement adhesion coefficient of the network is more than 95%.(2 Aiming at the defect that the existing minimum safe distance model can not be applied to some lane changing scenes,an emergency lane changing obstacle avoidance decision based on Minimum Distance and Collision Detection(MDCD algorithm is designed,which can make safe and effective decisions in all lane changing scenes.Firstly,the critical safety distance model of emergency braking considering road adhesion conditions is established;When the emergency braking can not effectively avoid obstacles,the emergency lane changing obstacle avoidance trajectory is planned based on the quintic polynomial,and the maximum lateral acceleration is limited by combining the road adhesion information and the point mass vehicle dynamics model;Finally,the MDCD algorithm is used to calculate the minimum distance between the self vehicle and each vehicle when changing lanes in theory,judge whether there is a collision,and limit the minimum distance in different scenes to ensure the safety of lane changing.After meeting the limit requirements,the emergency lane changing obstacle avoidance operation is carried out.(3 Aiming at the problems of low degree of freedom and low tracking accuracy of monorail vehicle dynamics model,a more accurate seven degree of freedom vehicle dynamics model is established.Based on this,a lane changing trajectory tracking controller of linear time varying model predictive control(LTV MPC is developed.The simulation results show that compared with the single track vehicle dynamics model(STVDM,the seven degree of freedom vehicle dynamics model established in this paper improves the trajectory tracking accuracy of the controller,and the tracking accuracy of the controller is greatly improved under the low adhesion road.(4 Three different simulation scenarios are designed based on MATLAB / Simulink Carsim joint simulation platform to verify the effectiveness of emergency lane change obstacle avoidance decision and the performance of trajectory tracking controller.The simulation results show that the system can make safe and effective obstacle avoidance decisions under the conditions of no car,fast car and slow car in the target lane;Moreover,the trajectory tracking controller can not only accurately track the lane changing trajectory on high,medium and low adhesion roads,but also has high real-time operation.