Multirobot Target Searches in Unknown Environments Via Waypoint Planning System

Target searches by multiple robots in unknown environments have been widely applied in practical applications, such as rescues in hazardous areas, environmental monitoring, and locating leak sources. However, in existing methods, it is difficult to find all the information-lack targets in the unknown environment efficiently. To fill this gap, we propose a multirobot waypoint planning system consisting of an area decomposition-based local waypoint planner and a submap-based global waypoint planner. The local waypoint planner generates efficient coverage waypoints for every robot and accelerates the acquisition rate of global information. In contrast, the global planner optimizes the planning of waypoints using global information and prevents the multiple robots from duplicating searches. The proposed planning system benefits from the two planners by simultaneously guaranteeing a high coverage rate and searching efficiency in unknown environments. The effectiveness and advantages of the planning system are verified through both simulations and experimental results.

I. Introduction

With decades of development in robot technology, autonomous robots have assisted (or replaced) humans to accomplish various tasks on many occasions [1], [2]. It is known that multirobot system is more flexible, robust, and parallel to perform complex tasks, which continues to attract the attention of scholars [3], [4], [5], [6]. Multirobot target searching is an important topic with many essential applications, such as search and rescue in hazardous areas, environmental monitoring [7], battlefield perception [8], and locating sources of gas, liquid leaks, and odor [9], [10].