Research on Path Tracking Control of Driverless Trucks

Abstract

AbstractPath tracking is a crucial function for achieving unmanned driving. This paper addresses the challenge of low tracking accuracy and poor stability in driverless trucks caused by uncertain model parameters and steady-state errors during path tracking. A linear quadratic regulator (LQR) controller optimization by an improved genetic algorithm has been designed. Firstly, the paper formulates the dynamic model of a two-degree-of-freedom vehicle as well as the model for tracking error. Subsequently, path tracking control is achieved through the utilization of feedforward control and LQR feedback control algorithms. Secondly, the weight coefficient of the LQR controller is enhanced through the utilization of an improved GA in order to boost the precision of path tracking. Ultimately, the devised controller undergoes simulation and validation in the TruckSim-Matlab/Simulink platform across diverse operational circumstances. The findings from the simulation demonstrate that the controller, optimized through improvements in the genetic algorithm, exhibits excellent tracking accuracy and stability.