Heterogeneous Mission Planning for Multiple UAV Formations via Metaheuristic Algorithms

Abstract

Unmanned aerial vehicle (UAV) formation has been widely applied in various aspects, both for military and civil purposes. In this paper, multiple UAV formations are deployed to perform different types of tasks, and the heterogeneous mission planning problem is studied. The problem is divided into two subproblems, i.e., task assignment and path planning. First, the task assignment is formulated into a combinatorial optimization problem. Different from the existing studies, appropriate UAVs need to be selected to build the formation and perform specific types of tasks. To obtain the task assignment scheme, a hybrid genetic and simulated annealing (HGSA) algorithm is proposed. In this algorithm, a feasible solution is guaranteed by the designed task-based strategy, and the diversity of solutions is ensured by the group-based solution remaining (GBSR) approach. In the path planning problem, an obstacle avoidance-enabled consensus (OAEC) algorithm is developed to form the UAV formation, which extends the application of the standard consensus algorithm. A multi-step particle swarm optimization (MPSO) algorithm is combined with the consensus algorithm to generate a path for the UAV formation to reach each task point. The conflict among multiple UAV formations is resolved by adjusting their departure time. Furthermore, the task assignment scheme is modified by utilizing real path information on the flying distance between the UAVs and the task points. The simulation results demonstrate that the HGSA algorithm can build the UAV formations and assign the tasks to them while satisfying all the complicated constraints. The advantages of the OAEC algorithm and the MPSO algorithm in path planning are verified by comparison with relevant algorithms.