Task Assignment of UAV Swarms Based on Auction Algorithm in Poor Communication Environments

Abstract

Poor communication environments always lead to unstable communication in unmanned aerial vehicle swarms. To solve the problem of task assignment in poor communication environments, this study proposed an information fusion strategy (IFS) based on information integrity and authenticity. The proposed IFS was embedded into the classical sequential and the Prim assignment and its generalization (G-Prim) decentralized task assignment algorithms, and these two improved variants with the proposed IFS were denoted as sequential auction with IFS (Seq-IFS) and G-Prim-IFS, respectively. The Bernoulli and Gilbert–Elliott models, which can model communication delay and packet loss, were adopted to describe unstable communication channels. A series of test instances with different swarm sizes and levels of communication channel reliability was used to test the performances of Seq-IFS and G-Prim-IFS in their original forms. Numerical experimental results demonstrated that the proposed Seq-IFS and G-Prim-IFS significantly outperformed their original versions in most test instances, particularly in cases with low communication environments.