Trajectory Tracking Control of Parking Automated Guided Vehicles Using Nonlinear Disturbance Observer-based Sliding Mode

Abstract

Automated valet parking systems based on parking automated guided vehicles (P-AGVs) are effective for improving parking convenience and increasing parking density. The ability of P-AGVs to move towards any position and attain any orientation simultaneously due to their mecanum wheels makes it convenient to transport vehicles in a parking lot. In this study, a nonlinear disturbance observer-based sliding mode controller for the trajectory tracking problem of a P-AGV is proposed. The kinematic and dynamic models for a P-AGV tracking trajectory are first analyzed in sequence and the influences of disturbing forces considered. Subsequently, a nonlinear disturbance observer (NDO) is designed to estimate the disturbing forces and torques generated by the caster wheels. Based on the designed NDO, a robust nonsingular terminal sliding-mode (NTSM) controller is used to track reference trajectories. The stabilities of the NDO and NDO-NTSM control systems are theoretically verified using their Lyapunov functions. Finally, simulations and experiments are performed to verify the effectiveness of the proposed control scheme. The experimental results show that the proposed NDO-NTSM controller can improve the trajectory tracking stability by 42–68% compared to a traditional NTSM controller. The NDO-based sliding mode controller for trajectory tracking proposed in this study can effectively reduce the impact of disturbances on trajectory tracking accuracy.

1. Introduction

In recent years, the rapid increase in car ownership in urban areas coupled with the small number of new parking spaces has made parking increasingly difficult [1]. Parking is one of the most challenging tasks in driving in which the driver has to control the steering and braking of the vehicle simultaneously while moving backward in a narrow space [2]. To overcome the difficulties associated with parking, smart parking systems have been installed in many lots [3–5], and some automakers have also installed automated parking systems [6] or automated valet parking (AVP) systems on vehicles [2, 7].

Differing from the abovementioned solutions, an automated guided vehicle (AGV)-based AVP system uses parking AGVs (P-AGVs) to automatically transport vehicles [8]. This allows omnidirectional movement to any location without requiring the vehicles to have automated driving capabilities or significant modifications to parking lots [8, 9].

P-AGVs have attracted considerable research attention owing to their remarkable advantages [10]. The technical solutions for vehicle exchange can be grouped into three types [11]. One type of P-AGV involves a platen [12]. The vehicle is first parked on the platen. The P-AGV beneath the platen then lifts the platen to move the vehicle. Another type of P-AGV similar to platen P-AGVs is the comb P-AGV, in which comb arms are used to lift the tires of vehicles and a set of matching comb racks is used to park the vehicles [13]. Comb P-AGVs differ from platen P-AGVs in that in the former, the comb arms can pass through a comb rack to lift the tires and the comb rack is not carried as the vehicle is transported [14]. A new type of P-AGV called clamping P-AGV was recently developed, in which tire-clamping arms automatically clamp and lift tires without the assistance of any other device [15, 16].