Vision-based coverage navigation for robot trash collection task

This paper presents a novel vision-based coverage navigation approach for trash collection task in which a robot has to search and collect trash and then deposit trash into trash bin. The proposed coverage navigation approach is composed of boustrophedon Path-Planner and simple-rule based Navigator. The former employs simple back-and-forth searching strategy to find pieces of litter. The later enables a robot moves to a destination from a given start position. The random Path-Planner, obstacle-following based Navigator, and genetic algorithm based Navigator are developed for comparing purpose. Simulation results showed that the proposed method has better performance in terms of path-length, number of robot´s turns, and percentage of large-cell overlap.