A Practical Framework for Cleaning Robots

Abstract

This paper presents an extension to the pattern-based genetic algorithm for multiple cleaning robots to achieve complete coverage path planning in an unknown environment. The extension is formulated in the form of a framework which consists of four phases. The phases are scouting, task distribution, cleaning, and confirmation. The scouting phase allows the robots to scout in the initially unknown floor plan. The task distribution phase distributes cleaning tasks to multiple robots. The cleaning phase uses the pattern-based genetic algorithm with an added function to cater to unforeseen obstacles. The confirmation phase recleans all the tiles. The performance of our proposed approach have been evaluated with six different floor plans through computer experiments. The cleaning phase performs better than the generic pattern-based genetic algorithm approach.