Intelligent Vehicle Path Tracking Control Based On Adaptive PP And MPC

The intelligent vehicle technology can realize autonomous vehicle operation,free the driver’s hands from the complex driving environment,reduce the frequency of traffic accidents and avoid traffic jams.Path tracking is an important link for intelligent vehicle to achieve their intelligent behavior.At the same time,accurate and stable tracking of the planned path is the basis for intelligent driving.The problem of intelligent vehicle path tracking control is studied in the paper.Firstly,two-wheel vehicle geometric steering model suitable for low-speed conditions is established.At the same time,two-degree-of-freedom vehicle dynamic model considering dynamic characteristics is also established.Based on the geometric steering model of the two-wheel vehicle,the pure pursuit algorithm is studied.Through simulation,the effect of the preview distance value in the pure pursuit algorithm on the path tracking performance is analyzed.The preview distance adaptive control strategy that comprehensively considers vehicle speed and expected road curvature factors is proposed,and the preview distance adaptive pure pursuit path tracking controller is designed.Based on Carsim/Simulink to build the co-simulation model,the simulation results show that the tracking performance of this method is better at low vehicle speed,but with the increase of vehicle speed,the defects of the two-wheel vehicle geometric steering model began to highlight,and the performance of vehicle path tracking is sharp descent.The vehicle suffered serious instability.Secondly,based on the two-degree-of-freedom vehicle dynamic model,the model predictive control algorithm is studied.The gaussian attenuation function is introduced into the model predictive algorithm to redistribute the tracking deviation weight value in the prediction time domain and the control increment weight value in the control time domain.Through simulation analysis,the range of the standard deviation of the gaussian function that has a great influence on the path tracking performance is determined.Genetic algorithm is used to optimize the gaussian function standard deviation.The weight matrix adaptive control strategy is proposed and the weight matrix adaptive model predictive path tracking controller is designed.Based on Carsim/Simulink to build the co-simulation model,the simulation results show that compared with the model predictive control method with fixed weight matrix,the weight matrix adaptive model predictive control method has better tracking accuracy under the premise of ensuring the tracking stability.Compared with the preview distance adaptive pure pursuit path tracking control algorithm,the tracking performance of the two control algorithms at low speed is not much different,but the tracking performance of the weight matrix adaptive model predictive control method at high speed is better.Finally,in order to further verify the effectiveness of the proposed preview distance adaptive pure pursuit control algorithm and weight matrix adaptive model predictive control algorithm in real controller.The hardware-in-the-loop test system is used to simulate and test the two proposed tracking control algorithms.The controlled object model is compiled and deployed to the NI real-time simulator through Veristand software.With the help of the D2 P platform,the two control algorithms in Simulink are flashed into the real controller.Hardware-in-the-loop tests are performed on the two control strategies designed.The results show that the weight matrix adaptive model predictive control algorithm and the preview distance adaptive pure pursuit control algorithm at low speed have good tracking performance.