Exploration in a boundary environment with unknown obstacles using Reaction-Diffusion equation on a graph

Abstract

In this paper, we propose a new exploration re-planning framework integrated with Reaction-Diffusion on a Graph (RDEG) for exploration in a boundary environment with unknown obstacles. In many cases, we have prior boundary information of an environment, utilizing this information has the advantage of creating more efficient coverage path planning. RDEG is an efficient method for coverage path planning utilizing prior boundary information. RDEG makes path planning with a good adaptation to a change of environment with less computational cost. RDEG utilizes the boundary information to make efficient path planning. However, the former method with RDEG is weak when dealing with an unknown obstacle. The integration of our proposed method removes this disadvantage with a new re-planning algorithm and a way to decide obstacle area using connected component labeling. Simulations are used to demonstrate our proposed method and compare it with other methods.