Research On Vehicle Path Tracking Control Based On Adaptive Horizon Variation

Rapid developed autonomous vehicles and related new technologies are supposed to be able to effectively alleviate the burden of intelligent traffic system.It emancipates drivers from the closed-loop control of "man-car-road" through the autonomous driving system.Path tracking control which aims to guarantee vehicles to track expected states or path is a key technology of motion control.It is mainly realized through lateral control and longitudinal control.The highly nonlinear characteristic of vehicle systems and complex coupling relationship between lateral and longitudinal expects much for tracking control algorithms.Aiming at the above problems,this paper focuses on the cooperative control of autonomous vehicles.The main work is as follows:(1)Model basic dynamics model for lateral or longitudinal tracking control.First,the three-degree-of-freedom nonlinear dynamics model is established for the basic predictive model of lateral tracking control.Secondly,the driving model and braking model of longitudinal tracking control are established.Finally,Pacejka’89 tire model was selected to study the variation characteristics of tire longitudinal and lateral forces,and then simplify the lateral dynamic model.(2)Lateral path tracking controller is designed based on model predictive control algorithm.Convert the dynamic model to predictive model,design the objective function and constraints to optimize system state and control variables.Consider design parameter—prediction horizon’s effect,7 groups of different speed and horizons are selected to conduct simulation and contrastive analysis to pick up a better parameter under different speed Then fit the functional relationship to design varying prediction horizon adaptive lateral path tracking controller which can real-time update prediction horizon.The simulation results shows the max lateral displacement error have decreased5% and 3.4%,the stability of vehicle has also been improved.(3)Layered longitudinal controller is taken to improve the effect of tracking speed.The upper controller is based on ACC control system,signals of front and manipulated vehicle are collected to calculate target acceleration,then send to lower controller to conduct logic switch between driving and braking modes.Inverse dynamic model is used to calculate throttle and brake pressure to change vehicle’s state.The co-simulation results show that minimum vehicle distance under adaptive following condition is 20 m,the maximum speed error under cruise control condition is 1.1km/h.(4)For decoupled independent lateral controller exists large speed tracking error,take speed as coupling point to design lateral and longitudinal cooperative control system to control front wheel angle,throttle and brake pressure.Based on Carsim and Simulink simulation platform to analysis tracking performance.