Research On Application Of Nonlinear MPC Algorithm In Intelligent Vehicle Path Tracking

Under the background of the continuous development of electronic,networked and intelligent vehicles,smart cars,as a kind of autonomous driving vehicle,have been the focus of attention for universities,enterprises and scientific research institutions.With the landing and trial operation of the domestic industrialization of intelligent driving,the government has put forward increasingly strict requirements on the safety and stability of smart cars.Path tracking technology,as a key technology in the intelligent vehicle system,is the basis and premise for ensuring its safe and stable driving.The development of high-performance path tracking technology has become an inevitable requirement for the sustainable development of the intelligent driving industry.With the development of computer technology,the model predictive control algorithm has gradually become an important research direction in the field of intelligent vehicle path tracking technology by virtue of its advantages in handling multivariate constraint problems.In order to meet the requirements of smart cars for the high real-time performance of the algorithm,most of the existing researches have built a linear MPC(Model Predictive Control)path tracker with high real-time performance by sacrificing the accuracy of the model,while the application of nonlinear model predictive control algorithm with higher model accuracy in path tracking is still relatively blank.In the general trend of system research and application from linear to nonlinear development,this paper proposes an application method of nonlinear MPC path tracking.Compared with linear MPC control,this method improves the stability and accuracy of high-speed path tracking.At the same time,it can adapt to various driving conditions and meet the high real-time requirements of the algorithm for path tracking.This paper first proposes the architecture design of the decoupling control layer based on the intelligent vehicle platform GE3,which has the characteristics of portability and scalability.Then,a more in-depth algorithm framework is designed for the horizontal control module in the architecture,and a lateral control module algorithm framework based on the vehicle state prediction server and with the deviation solver,data driver and MPC solver as the core is proposed.In order to improve the real-time performance of the algorithm,the second-order Runge-Kutta method is used to directly discretize the derived nonlinear prediction model,so as to solve the vehicle state matrix required by the model prediction,and then jointly construct the safety constraints and the objective function In order to establish the nonlinear planning problem of path tracking,the problem is finally solved and calculated by the SQP algorithm.In addition,in order to reduce the calculation amount of the algorithm,this paper also proposes a method for solving lateral deviation based on the integrated design of planning and tracking.In order to analyze the influence of algorithm parameters on the path tracking effect,an algorithm simulation platform based on Simulink was built.Based on the simulation results,the effects of the predicted time domain,vehicle speed,and weight parameter values on the path tracking performance of the algorithm are analyzed.It is established that optimizing the weight parameters is the key to improving the performance of the tracker.Finally,the Simulink genetic algorithm is used to optimize the weight parameter values at different speeds.The tracking performance of the unoptimized parameters is analyzed and compared to verify the effectiveness of the parameter optimization results.In order to further verify the performance of the algorithm,a hardware in the loop simulation platform based on prescan and ROS is built.By writing the C + + algorithm node program,the performance of the algorithm in the path tracking is tested,and the effectiveness of the NMPC control algorithm is verified.At the same time,in order to further verify the control effect of the algorithm on the vehicle in the real environment,this chapter carries out a real vehicle test based on the refitted intelligent vehicle platform,and tests the effect of the algorithm on the path tracking under the single lane change condition,avoiding accident vehicle conditions and the high curvature road condition respectively.The results of real vehicle test show that the algorithm can adapt to various road conditions,and its tracking accuracy,stability and real-time performance can meet the needs of practical application at a speed of up to 75 km/h?...