Road-Curvature-Range-Dependent Path Following Controller Design for Autonomous Ground Vehicles Subject to Stochastic Delays

• Published in: IEEE transactions on intelligent transportation systems Vol. 23; no. 10; pp. 17440 - 17450
• Main Authors: Shi, Qian, Zhang, Hui
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Actuators; Bicycles; Constraint modelling; Control systems design; Controllers; Curvature; Delays; Dynamic stability; Feedback control; grey wolf optimizer (GWO); Liapunov functions; Mathematical models; Model forms; Optimization; Output feedback; Path following; Proportional integral derivative; road-curvature-range-dependent PID controller; Roads; Stability criteria; static output feedback (SOF); Tracking errors; Trajectory planning; Tuning; Unmanned ground vehicles; Vehicle dynamics
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2022.3157059

In this paper, we investigate the PID controller design problem of path following for an autonomous ground vehicle (AGV). Firstly, a bicycle model is adopted and a vehicle offset model from the target path is integrated to the bicycle model. The PID controller considers the vehicle longitudinal speed variation for adaption to the road curvature and the stochastic delay induced by the network communication. This is realized by transforming the tuning problem of proportional-integral-derivative (PID) gains for path following into a design problem of a static-output-feedback (SOF) controller for a time-delayed linear parameter varying (LPV) model form. A sufficient condition is adopted to guarantee the stability of the closed-loop system. In order to achieve better tracking performance, we propose a strategy in which the PID gains are piecewise constant and are dependent on the road-curvature ranges. The stability of the switched system is guaranteed via the common Lyapunov function method. Grey wolf optimizer (GWO) is employed to solve the optimization problem with maximum absolute tracking error as the optimization objective and stability condition and actuator dynamics as constraints. Both simulation results based on the CarSim-Simulink joint platform and hardware-in-loop experiment results are used to verify the effectiveness of the proposed control strategy.