Research On Lane-changing Trajectory Planning And Tracking Control For Distributed Drive Intelligent Electric Vehicle

The global automobile industry is facing a major change that has not been seen in a century.Outline of the 14th Five-Year Plan for National Economic and Social Development and Vision 2035 pointed out that the industrial transformation characterized by the mutual promotion of electrification and intelligence should be actively promoted.The electrification,intelligence and networking of automobiles have become a major national development strategy.The distributed drive electric vehicle with in-wheel motors highly integrates the driving,braking and steering systems,which realizes the high integration of the chassis system.The redundant characteristics of multiple actuators endow the vehicle with better motion potential.Therefore,it is considered by the industry as the special chassis of electric vehicles and the best carrier to ensure high safety and efficiency.At this stage,lane-changing trajectory planning and tracking control are conservatively designed based on empirical values,and there are problems of insufficient safety in multi-vehicle traffic scenarios.Electrification and intelligence are still in an independent industrial layout.At the same time,distributed drive electric vehicles also put forward new challenges to the intelligent driving architecture,and the development of intelligent automobiles is still a long way to go.The driving stability and motion reliability of intelligent vehicles are important criteria.How to reveal the dynamic mechanism of chassis motion stability region,match the intelligent driving platform and chassis system,and broaden the kinematic interface of active safety to dynamic interface are key problems to be solved urgently.In this paper,the distributed drive intelligent electric vehicle is taken as the research object,and high-speed,emergency collision avoidance and other multi-vehicle traffic scenarios are taken as the background.The full-state estimation,stability region analysis,lane-changing trajectory planning and trajectory tracking control are studied in depth.The main contents are as follows:（1）It is rather difficult to measure side slip angle and tire lateral force of the distributed drive electric vehicles.Considering the unmodeled dynamic characteristics of the system,model parameter perturbation,system process noise and measurement noise,a Nonlinear Robust Fusion Estimation method is proposed.The vehicle longitudinal dynamic equation considering longitudinal air resistance and tire rolling resistance is established,and a forgetting factor recursive least square is proposed to accurately estimate the mass.A three-degree-of-freedom dynamic model is established,and a semi-empirical magic tire model that can characterize the instantaneous mechanical characteristics of tires is introduced.Under the framework of Robust Cubature Kalman filter,a vehicle dynamics information fusion model is constructed to estimate side slip angle and tire lateral force in real time.The estimation error minimization in the background of the maximum value is embedded into the Cubature Kalman filter.Under the premise that Q_k、R_k and P₀ are unknown,Robust Cubature Kalman filter minimizes the influence of uncertainty of W_k、V_k and X₀ on the accuracy of estimation results.The experiments show that the NRFE algorithm is verified to have better accuracy and robustness compared to RCKF and CKF by the mean absolute deviation and the root mean square error.（2）To address the problem that multi-actuator can broaden the stability region of electric vehicles while the instability regions are difficult to characterize,and the dynamic of chassis system is important,a vehicle stability region boundary equation based on a physical theory model is proposed.The bicycle model and the four-wheel nonlinear dynamic model are established.The vehicle model which can accurately describe the nonlinear characteristics is determined by the test vehicle.Based on the phase plane,the evolution of the equilibrium point,saddle point and track line are revealed.According to the physical theoretical model,the expressions of the stability region boundary are derived.Through the boundary equation,the expression of the conservative stability region can be derived.According to the problems that are difficult to reveal the dynamic of nonlinear system,a new method about expansion stability region is proposed,and the stability regions are quantitatively analyzed.The range of direct yaw moment is calculated according to the tire adhesion margin.The mechanism of direct yaw moment intervention on the nonlinear system under different torque distribution methods is analyzed,and the boundary of the stability region is determined.（3）Aiming at the limitations of the autonomous lane-changing system in high-speed,emergency collision avoidance and other multi-vehicle traffic scenarios,a lane-changing trajectory planning and tracking control method based on stability region is proposed.The optimal boundary value problem is constructed,and the cost function index optimization problem is solved according to Pontryagin’s Minimum Principle.A series of unconstrained generalized lane-changing trajectory clusters based on quintic polynomials are calculated.Considering the stability region,environmental constraints and road boundaries,the global feasible region is divided.A comprehensive evaluation index system is established.Based on the TOPSIS algorithm,a multi-objective comprehensive evaluation is carried out to obtain the optimal trajectory.The three-degree-of-freedom dynamic model and trajectory tracking error model are established.The state equation,performance index function and quadratic programming problem of the prediction model are constructed.The LPV-MPC trajectory tracking controller is designed based on the stability region.The simulation results show that the vehicle realizes the optimal trajectory planning and tracking control,which improves the robustness,stability and feasibility of the system compared with the traditional algorithm.（4）The intelligent driving system of distributed drive electric vehicles is developed.The chassis system and communication network architecture are designed based on production vehicles.The vehicle control software is developed based on V model process.The model is verified,optimized,tested and integrated.The control system is developed,including the design of steer-by-wire system,speed control system and pedal curve.The stability and reliability of the chassis system are verified by the test vehicle.The intelligent driving platform is designed,and centimeter-level positioning is realized through GNSS/INS.Based on the ROS system,the upper control system and the control software system are completed.The system test based on the developed distributed drive intelligent electric vehicle verifies the tracking control of different global trajectories and lane-changing trajectories,and improves the real-time,accuracy and robustness of vehicle trajectory tracking.