Trajectory Tracking and Obstacle Avoidance for Wheeled Mobile Robots Based on EMPC With an Adaptive Prediction Horizon

This article develops an event-triggered model-predictive control (EMPC) strategy to realize trajectory tracking and obstacle avoidance for a wheeled mobile robot (WMR) subject to input constraints and external disturbances. In the EMPC strategy, a potential field is introduced in the cost function to guarantee a smooth path for the WMR. An event-triggered mechanism is designed to reduce the computational load of solving an optimal control problem (OCP). Moreover, an adaptive prediction horizon is utilized to further achieve computation reduction. Both recursive feasibility of the OCP and practical stability of the resulting closed-loop system are analyzed for the WMR with the input constraints and the external disturbances. Simulation results are provided to demonstrate the effectiveness and superiority of the proposed EMPC strategy.

I. Introduction

In recent years, wheeled mobile robots (WMRs) have been exploited in a wide range of fields, such as military, industry, and agriculture [1], [2]. It is noted that trajectory tracking and obstacle avoidance control is the key to accomplish missions for the WMRs [3]. However, the actual tracking performance is inevitably deteriorated by external disturbances [4]. To this end, several control strategies on trajectory tracking have been constantly proposed for the WMRs with external disturbances. In [5], robust adaptive tracking control has been investigated via a Gaussian wavelet network for a WMR with wheel slips. In [6], a disturbance-observer-based approach has been employed for the WMR to reject external disturbances and guarantee desired tracking performance. In addition, some intelligent control technologies have also been developed for the trajectory tracking and achieved satisfactory results [7], [8]. Note that the WMRs usually work in complex and changing environments; hence, they unavoidably encounter obstacles when tracking a given reference trajectory [9]. In [10], a nonlinear control scheme has been designed for the WMR to solve a trajectory tracking and obstacle avoidance problem. However, an obstacle avoidance path generated by the nonlinear control scheme is not always smooth. In [11], a model-predictive control (MPC) method integrated with a potential field has been utilized to avoid obstacles smoothly and guide the WMR to a goal successfully. Moreover, a control Lyapunov-barrier function-based MPC method has been developed in [12], which provided a satisfactory solution for safe obstacle avoidance. A multiple obstacle case is also able to handle by utilizing the control Lyapunov-barrier function, more technical details have been sorted out in [13]–[15] and the references therein. Furthermore, a combined control strategy has been proposed for an automated guided vehicle based on a path planner and an MPC [16]. It is interesting to design an MPC strategy to realize accurate trajectory tracking and smooth obstacle avoidance for the WMR subject to external disturbances.