Path Planning for Multiple UAVs Using Artificial Intelligence Under Changing Environment

초록

If the disaster like a forest fire or marine distress has occurred, Unmanned Aerial Vehicles (UAVs) can be utilized for fast searching in the area. Although an expensive and powerful UAV can accurately explore a wide area, it can't search larger area faster than multiple UAVs. The easiest meth-od for operating multiple UAVs in area search problem is to divide the search areas and assign them to each UAVs. With this area allocation method, distributed control that each UAV control themselves is appropriate. In order to perform precise area searching tasks in a short time, moving all areas without duplication is needed, which directly related to how efficient Complete Coverage Path planning (CPP) is. This paper suggests the CPP algorithm by using deep Q-learning in an unpredictable environment. The performance of the path planning is compared to the path planed with the genetic algorithm (GA).