Autonomous ground cover clearing utilizing a robotic wheel loader

Abstract

The ongoing need to remove ground covering such as sand, dirt and snow poses dangers to workers, from fatigue and repetitive work injures, to direct harm in hazardous locations. To address these challenges a robotic system capable of autonomous clearing of non-compacted material is presented, focusing on detection, clearing strategy and clearing validation. The robotic design is based on an articulating wheel loader, electrically driven with hydraulic articulation. The sensor suite comprises a depth camera used for Simultaneous Localization and Mapping (SLAM) and material detection. Potentiometers for joint positioning, motor encoder for distance measurements and an ultrasonic rear sensor for obstruction detection are implemented on the loader. To detect ground cover, pixel locations inline with the bucket provide depth information and subsequent ground cover validation. An overlapping path strategy is used to minimize material spillage based on bucket volume control. Testing of the system showed it was successful in ground cover detection, material removal and minimizing material spillage.