Research on Collaborative Task Allocation of Heterogeneous UAVs With Complex Constraints

This paper addresses the multi-task allocation problem with complex constraints in the multi-UAV system by presenting an extension to consensus-based bundle algorithm(CBBA). The presented algorithm termed dynamic clustering consensus-based bundle algorithm(DCCBBA) provides an improved bundle construction, a novel consensus strategy and a systematic way of grouping. DCCBBA features different consensus methods for inner conflict resolution and outer conflict resolution. The inner conflict resolution aims at handling the violated task assignment that belongs to the same cluster, while outer conflict resolution guarantees conflict-free assignments in the overall UAVs. The proposed DCCBBA coupled with Pythagorean Hodograph curves can significantly reduce the number of communication events required to achieve consensus while preserving the robust convergence property of the baseline CBBA. Numerical simulations on various cooperative reconnaissance and strike tasks demonstrate the reduction of communication events compared with the baseline CBBA, the cluster-formed CBBA, the grouped CBBA and Two-layer CBBA.