A Complete Coverage Path Planning Algorithm for Cleaning Robots Based on the Distance Transform Algorithm and the Rolling Window Approach in Dynamic Environments

Abstract

This paper presents an algorithm of complete coverage path planning (CCPP) for cleaning robots in dynamic environments based on the rolling window approach and the distance transform algorithm. Before the coverage task, the robot models the static environment with a global grid map as priori knowledge. In the process of coverage, a rolling window is created, corresponding to which a local grid map is extracted from the global map. The robot uses on-board sensors to detect local environments (including dynamic obstacles) and updates the local grid map, based on which the distance transform algorithm is adopted to produce a covering path in the rolling window. After that, the robot will perform the CCPP task by moving along with the planned path. In order to deal with dynamic obstacles, the robot has to update the local grid map and plan a coverage path in real time. The updating and planning procedures will be carried out repeatedly once the robot has covered a cell, and will be stopped when the robot has covered all the areas. To verify the proposed method, we have compared it with the method which combines with the biologically inspired neural networks and rolling path planning. Simulation results show that the proposed method can make the robot cover the entire workspace with lower repetition rate and shorter trajectory length in complicated dynamic environments.