MPC and PSO Based Control Methodology for Path Tracking of 4WS4WD Vehicles

Qifan Tan,Penglei Dai,Zhihao Zhang,Jay Katupitiya

doi:10.3390/app8061000

Abstract

Four wheel steering and four wheel drive (4WS4WD) vehicles are over-actuated systems with superior performance. Considering the control problem caused by the system nonlinearity and over-actuated characteristics of the 4WS4WD vehicle, this paper presents two methods to enable a 4WS4WD vehicle to accurately follow a predefined path as well as its reference trajectories including velocity and acceleration profiles. The methodologies are based on model predictive control (MPC) and particle swarm optimization (PSO), respectively. The MPC method generates the virtual inputs in the upper controller and then allocates the actual inputs in the lower controller using sequential quadratic programming (SQP), whereas the PSO method is proposed as a fully optimization based method for comparison. Both methods achieve optimization of the steering angles and wheel forces for each of four independent wheels simultaneously in real time. Simulation results achieved by two different controllers in following the reference path with varying disturbances are presented. Discussion about two methodologies is provided based on their theoretical analysis and simulation results.

Highlights

With the development of Autonomous Ground Vehicles (AGVs) in the last few decades, the demand for accuracy, maneuverability and controllability in vehicle’s navigation is ever increasing.For example, an AGV may be required to follow a path accurately under unstructured and uneven terrain conditions, where a significant amount of wheel slip and unpredictable disturbance forces occur at the vehicle’s wheels
The main challenge in the control of 4WS4WD control is the number of control inputs, which results in an over-actuated system, where only three outputs including its degree of freedom (DOF) in the longitudinal, lateral and angular directions of the vehicle are concerned
This paper presented two control methodologies applicable to four wheel steering and four wheel drive vehicle systems to track paths accurately

Summary

Introduction

With the development of Autonomous Ground Vehicles (AGVs) in the last few decades, the demand for accuracy, maneuverability and controllability in vehicle’s navigation is ever increasing.For example, an AGV may be required to follow a path accurately under unstructured and uneven terrain conditions, where a significant amount of wheel slip and unpredictable disturbance forces occur at the vehicle’s wheels. With the development of Autonomous Ground Vehicles (AGVs) in the last few decades, the demand for accuracy, maneuverability and controllability in vehicle’s navigation is ever increasing. The 4WS4WD vehicle, with four wheels that can be steered and driven independently, is a revolutionary platform that has great potential to perform high maneuverability and flexibility in harsh environments. The main challenge in the control of 4WS4WD control is the number of control inputs (four steering angles and four drive torques), which results in an over-actuated system, where only three outputs including its degree of freedom (DOF) in the longitudinal, lateral and angular directions of the vehicle are concerned. How to allocate all eight control inputs to achieve high path following performance has not yet been effectively solved. The control allocation can be treated as an optimization problem

Results

Discussion

Conclusion