Development of Intellectual Online Coverage Path Planning Algorithm for Cleaning Robot
- 원문 URL
The complete coverage problem of service robots is applied to a wide range of applications, such as floor cleaning, lawn mowing, mine hunting, and harvesting. This problem can be performed by either single- or multi-robot systems and is considered an optimal criterion in terms of the coverage rate and time. At a high level, the complete coverage task can be solved by either offline or online method. The offline method is to determine a coverage path for each robot based on a known map of the workspace so that the union paths cover the entire workspace. Then, the robots follow the determined paths to perform the coverage task. By contrast, in the online method, the robots accomplish the coverage task without any prior knowledge about the workspace and thus must construct their coverage paths step by step during the execution of the coverage algorithm. In real-world scenarios, where the map of a workspace is unknown to the robots, online complete coverage is preferred. Thus, the development of this method should be given considerable attention. From this point of view, we focus on developing a technique for the online complete coverage task of the single-robot systems based on spiral motion, intellectual boustrophedon and A* search, called SIBA algorithm, to work in unknown environments. SIBA algorithm mimics human behavior by tuning boustrophedon algorithm to clean regions from left to right sequentially. Computer simulations show better performances from the 4 viewpoints compared to other algorithm: The number of sweeping regions, path length, shorter time and better moving style like human.