Research On Vehicle State Parameter Estimation Algorithm Based On Combination Of Kinematics And Dynamics

The estimation of vehicle driving state parameters is the foundation for achieving vehicle control,navigation,and assisted driving functions.Due to the limitations of a single sensor itself,some vehicle state parameters cannot be directly measured.Therefore,in this paper,the vehicle driving data is collected through on-board sensors,and the longitudinal speed,sideslip angle and yaw rate of the vehicle are accurately estimated by combining the vehicle kinematics and dynamics models.The specific research contents are as follows:First of all,accurate collection of vehicle driving data is the basis for estimating vehicle state parameters.However,a single sensor is greatly affected by noise when collecting data,which can lead to error accumulation and data divergence,making it difficult to use for a long time.In response to this issue,an INS/GPS loose combination navigation system based on the EKF fusion algorithm was established.The GPS output pose information was used to compensate for the error accumulation generated by the INS during data collection.Subsequently,Matlab programming and simulation were conducted to verify the high accuracy of the fusion algorithm,which achieved accurate estimation of vehicle speed,acceleration,and pose information,and provided guidance for subsequent vehicle centroid sideslip angle,longitudinal velocity Accurate estimation of yaw rate provides assurance.Then,the accurate estimation of vehicle state parameters needs to make full use of the information of dynamics and kinematics models.Aiming at the problem of poor estimation accuracy of ordinary vehicle dynamics models,an AUKF vehicle state parameter estimation algorithm based on forgetting factor is constructed.Based on the principle of different tire loads during vehicle operation,the recursive least squares method is introduced to obtain real-time tire lateral stiffness,which improves the measurement accuracy of the dynamic model.Finally,through Carsim/Simulink joint simulation and comparative experiments,it was proven that the above model algorithms have high accuracy in estimating the information of center of mass sideslip angle,yaw rate,and longitudinal acceleration;An error correction equation is designed for the kinematics model to reduce the error accumulation caused by long time integration and improve the accuracy of the integration method.according to the distribution characteristics of the sideslip angle of the center of mass estimated by the dynamic model and the kinematics model,a fusion algorithm of the sideslip angle of the center of mass based on the two-dimensional extension theory is designed.With the lateral velocity and lateral acceleration as the dependent variables,the weight coefficient of the fusion model is modified,and the proportion of the estimated values of the above two model algorithms is allocated.The fusion algorithm has been verified to have good robustness and tracking performance under different operating conditions through Carsim/Simulink joint simulation comparison experiments and real vehicle experiments.