Trajectory tracking switching control system for autonomous crawler dump under varying ground condition

Abstract

This study proposes an autonomous crawler dump trajectory tracking control system considering varying ground surfaces. The proposed system comprises three-mode proportional–integral controllers, a switching mechanism, and a ground condition estimator. The need for controller switching is clarified through model identification of a crawler dump using the input and output (sensor) signals measured from Kajima’s experimental field with hard and soft surfaces. The features in the measured z-axial acceleration signal help construct the ground condition estimator using a support vector machine. The resulting trajectory tracking errors were evaluated numerically and statistically to verify system effectiveness. Moreover, the superior performance of the three-mode control system was highlighted by comparing the numerical evaluation results with that of a two-mode switching control system. Furthermore, the proposed simulation-model-based control system design can be extended to a multiple-mode switching control system considering several ground conditions for autonomous crawler dumps.