EA-OSPGB: Multiple robots dynamic online algorithm for solving full coverage path planning of multiple robots in unknown terrain environments

Abstract

To address the issue of high energy consumption resulting from poor synergy among multiple robots during full-coverage dynamic online planning in unknown terrain, this paper proposes a multi-robot coverage algorithm guided by the energy activity (EA) function. Additionally, a backtracking mechanism based on the terrain environment is incorporated. First, occupancy grid is employed to represent the area to be covered, with the local raster activity value function guiding the coverage of the working environment. Next, a terrain-based backtracking mechanism is incorporated into the algorithm to facilitate online collaboration among the robots and help them escape “dead zones,” thereby preventing conflicts in backtracking areas and reducing the likelihood of lengthy backtracking paths. Finally, by simulating various scenarios that a cleaning robot may encounter in an unknown terrain environment, we compared the results with those of other algorithms and with scenarios that did not consider terrain factors. The experimental results demonstrate that accounting for terrain is more effective in reducing the robot’s energy consumption. The experiments conducted in different situations highlight the benefits of considering terrain factors. Specifically, the average path length and the number of turns were reduced by 5.2 % and 30.5 % compared to the BOB algorithm, and by 3.1 % and 19.3 % compared to the ${\epsilon }^{★}$ algorithm. Thus, the feasibility and effectiveness of the proposed algorithm are confirmed.