A Two-Stage Method for Target Searching in the Path Planning for Mobile Robots

Abstract

The path planning for target searching in mobile robots is critical for many applications, such as warehouse inspection and caring and surveillance for elderly people in the family scene. To ensure visual complete coverage from the camera equipped in robots is one of the most challenging tasks. To tackle this issue, we propose a two-stage optimization model to efficiently obtain an approximate optimal solution. In this model, we first develop a method to determine the key locations for visual complete coverage of a two-dimensional grid map, which is constructed by drawing lessons from the method of corner detection in the image processing. Then, we design a planning problem for searching the shortest path that passes all key locations considering the frequency of target occurrence. The testing results show that the proposed algorithm can achieve the significantly shorter search path length and the shorter target search time than the current Rule-based Algorithm and Genetic Algorithm (GA) in various simulation cases. Furthermore, the results show that the improved optimization algorithm with the priori known frequency of occurrence of the target can further improve the searching with shorter searching time. We also set up a test in a real environment to verify the feasibility of our algorithm.