Cooperative task planning for multiple unmanned aerial vehicles using a genetic algorithm

Abstract

This paper addresses the mission planning issues for guiding a group of unmanned aerial vehicles to carry out a series of tasks, namely classification, attack, and verification, against multiple targets. The flying space is constrained with the presence of flight prohibit zones (FPZs) and enemy radar sites. The solution space for task assignment and sequencing is modelled with a graph representation. With a path formation based on Dubins vehicle paths, a genetic algorithm has been developed for finding the optimal solution from the graph to achieve the following goals: (1) completion of the three tasks on each target, (2) avoidance of FPZs, (3) low level of exposure to enemy radar detection, and (4) short overall flying path length. A case study is presented to demonstrate the effectiveness of the proposed method.