Research On Trajectory Planning And Path Tracking Control Strategy For Intelligent Vehicle Emergency Collision Avoidance

The intelligent vehicle is the product of the integration and development of advanced sensing and artificial intelligence and other emerging technologies.It is a new generation of vehicles with an automatic driving function.It is also the key to solve the problems of traffic accidents,low vehicle utilization and traffic congestion.The development of intelligent vehicles is of great significance to accelerate the transformation and upgrading of China's automobile industry.The research on the emergency collision avoidance of intelligent vehicles plays an important role in promoting the application of intelligent driving technology and improving driving safety.At the same time,it has positive significance for the development of vehicle motion planning and vehicle motion control technologies.This dissertation takes the intelligent vehicle as the research object and deeply studies the collision avoidance trajectory planning strategy,path-tracking control system modelling,and path tracking control strategy as well as its optimization strategy.The research on the trajectory planning strategy of the intelligent vehicle for emergency avoidance of the vehicle that enters the ego lane from the opposite direction is carried out.A trajectory prediction method for opposite vehicles and the trigger conditions for the trajectory planning strategy of ego vehicle is studied;the candidate trajectory planning scheme based on single-lane space is clarified;a path planning algorithm based on model-predictive planning method,a speed planning algorithm based on quintic polynomial with a “minimum speed” cost function and a target state determining method based on the BP neural networks are designed,and the collision avoidance trajectories are generated.A hierarchical collision checking algorithm and a vehicle collision type identification method are proposed.A collision severity prediction algorithm based on delta-v and vehicle collision type is proposed,and a multi-objective evaluation function for collision-free candidate trajectory is designed,then the collision avoidance trajectory with optimal performance for the intelligent vehicle is determined.The state-space model of the intelligent vehicle path tracking control system is established.The model is composed of a two-degree-of-freedom vehicle model,a tire lateral force calculation model,and a path tracking error model.A lateral dynamic model for the intelligent vehicle is designed,and a numerical calculation method for the desired steering angle based on the Fiala tire model and the numerical look-up table of front tire lateral force is studied.The skills of improving the modelling accuracy of tire nonlinearity are studied,while a rear tire lateral force linearization method for the Model Predictive Control(MPC)algorithm is proposed based on two-point affine;the path tracking error model based on the yaw angle deviation is established.The establishment of the state-space model of the intelligent vehicle path tracking control system is completed.The composite tracking error model based on the vehicle heading deviation is studied,and the path tracking control strategy for emergency collision avoidance is established.The predictive model and cost function of the MPC algorithm,as well as the MPC path tracking control algorithm,are established.The influence mechanism of the tracking error model on vehicle control performance is studied,and a composite tracking error model based on the vehicle heading deviation is proposed,which could achieve a good tracking accuracy under steady and transient steering conditions.The research on the characteristic parameters of the vehicle steering state is carried out,and the steering state identification algorithm based on the fuzzy logic and weighted average method is proposed;the driving stability constraints are studied,and the path tracking control strategy based on the composite tracking error model for emergency collision avoidance is completed.The research on robust optimization of path tracking control strategy is carried out,and a robust control strategy based on Tube Invariant Set Robust MPC(Tube-RMPC)for the intelligent vehicle path tracking is proposed.The typical uncertainties such as strong side wind in the intelligent vehicle driving environment are analyzed,and the control system model with uncertainties is established.By analyzing the requirements of Tube calculation methods of the path tracking control system,a novel calculation method based on the control matrixes polyhedron segmentation is proposed.By combining the off-line convex hull and the N-steps reachable sets,the Tube sequences are obtained.The closed-loop feedback gain of Tube-RMPC,the terminal cost and terminal invariant set,the admissible sets of the nominal system are calculated.The robust path tracking control strategy for the intelligent vehicle based on Tube-RMPC is established.Based on the Car Sim and Simulink simulation platform,the simulation tests of intelligent vehicle emergency collision avoidance trajectory planning and path tracking are carried out.The simulation results show that in the emergency collision avoidance condition with the vehicle approaching the handling limits,the proposed control strategy based on the composite tracking error model could maintain the lateral deviation within 0.1m.And compared with the widely-used control strategy based on the yaw angle deviation tracking error model,the proposed control strategy reduces the root mean square of the path tracking error by 28.6%.The proposed Tube-RMPC based robust path tracking control strategy could maintain the lateral deviation within 0.2m under the disturbance of 0.3 road friction uncertainty and the 25 m/s strong side wind in the emergency collision avoidance condition,which indicates that the robust strategy could keep the intelligent vehicle stable under uncertain conditions.The effectiveness of the designed trigger conditions,collision detection algorithm,collision severity prediction function and multi-objective comprehensive evaluation function for candidate trajectory has been verified;and the simulation results show that the proposed emergency collision avoidance trajectory planning strategy can make the intelligent vehicle achieve collision avoidance with a larger safety margin or achieve collision mitigation with a smaller predicted collision severity in the scenario of the opposite vehicle encroaches into the ego vehicle' lane.There are 91 figures,13 tables and 200 references in this dissertation.