State Estimation And Torque Allocation Strategy For Distributed Drive Electric Vehicles

With the transformation of energy stratagem,the share of electric vehicles has increased year by year in China.The research on electric vehicle control is becoming a new hot spot in the field of vehicle control,especially vehicles equipped with drive-bywire technology.In this paper,electric vehicles driven by in-wheel motor is taken as the research object.Based on vehicle dynamics,Kalman filter and model predictive control theory,this paper presents system modeling,parameter estimation and driving torque calculation.A Matlab/Simulink-Car Sim co-simulation platform is used for simulation verification,and the reliability and stability verification of the algorithm is completed through the vehicle-grade electric drive platform built by ourselves.The main research contents of this paper are as follows:Firstly,according to vehicle system architecture,the algorithm structure of vehicle state estimation and torque allocation control is analyzed,and the selection of sensors and actuators is completed.The E-E architecture of platform is designed,using CAN communication network and Ethernet to exchange data between computing units.The functional verification of each part is carried out for building this vehicle-grade electric drive platform.In order to solve the difficult acquisition of vehicle state,estimator based on RLS,UKF and EKF is proposed for tire cornering stiffness,motion state and position,combining with magic formula tire model and 3-DOF dynamic model.Especially when GPS/GNSS fails,a redundant longitudinal vehicle speed estimation method using wheel speed sensors and IMU is proposed to obtain vehicle’s longitudinal speed.The simulation results show that the tire cornering stiffness estimator can provide a more accurate tire cornering force estimation online according to the axle load transfer in DLC tests at different vehicle speeds.The UKF motion state estimator can obtain the accurate vehicle longitudinal speed,yaw rate and sideslip angle.The EKF position state estimator integrates GPS and IMU data to provide high-frequency and accurate information to meet the needs of automatic driving systems.The EKF longitudinal vehicle speed redundancy estimator can effectively eliminate the influence of tire slippage and IMU integral drift,and acquire accurate longitudinal vehicle speed when the positioning system fails.Aiming at the requirements of carrying passengers and loads,a joint control goal is proposed,considering vehicle handling stability and wheel slippage improvement.Four wheel driving torques are output to actuator.They are transferred to torque increment by error state space construction using MPC theory.The output is calculated via the solution of QP problem.The simulation results show that the proposed torque optimal distribution strategy can reduce the steering stability parameters including sideslip angle and yaw rate,and effectively reduce the slip rate of each wheel during the steady and transient driving condition.In addition,the driving torque of each wheel is also reduced to varying degrees,which improves the driving economy of whole vehicle.Finally,the real vehicle test was completed in a closed park with good friction condition.Under the medium and low speed DLC and the comprehensive driving conditions,the results of state estimator and torque distribution strategy are consistent with simulation.The validity,reliability and stability of algorithms above get verified in test vehicle.