Autonomous Drone Path Scheduling and Management Strategy with Multi-agent Decision Support and Coordination

Abstract

In a harsh and isolated environment with limited communication and human interaction, e.g., Antarctica, multiple unmanned vehicles or multiple drone systems can use fuel/battery faster than normal conditions. This has led to many drone failures. Preservation of fuel level becomes an essential factor for the management of drones while performing the required set of tasks within an acceptable time. The uncertainty of the environment further challenges the autonomous task allocation and path planning for drones. The existing literature algorithm was not tested or used in these severe conditions, which creates the major gaps in the autonomous drone’s management. This paper proposes a CBS-LSA algorithm, which is a combined algorithm of conflict-based search and linear sum assignment. This algorithm considers the drones’ fuel consumption before allocating the tasks uniformly and proportionately to all the available agents whenever the new sets of tasks were generated randomly. It can proportionately allocate tasks to drones, even if the number of drones is lesser than the number of tasks while maintaining the optimal task allocation process. Moreover, the algorithm optimizes the fuel consumption of the drones by allocating them to the nearest available agents and providing the best possible path for the agents that avoid static obstacles such as mountains and terrains.