Adaptive Assignment Re-Consensus in Communication-Constrained Environments

When employing swarms of unmanned aerial vehicles (UAVs) in communication-constrained environments, it is of vital importance to coordinate their actions in cooperation despite sparse and unreliable communication channels. This paper proposes an adaptive dual-phased threshold-based assignment scheme for robust coordination under lossy communication. The assignment scheme is inspired by features in handshake protocols, using records upon failed communications to keep track of swarm mates and consensus rates. Resendings of vital information pieces within core steps of assignment negotiation are arranged to increase consensus rates above the threshold. The resendings are constrained by switching criteria designed to balance between information integrity and assignment timeliness. The overall scheme is termed Robust Assignment under Lossy Communication (RALC). The proposed RALC is evaluated at various levels of communication reliability, using the Bernoulli model and the Gilbert-Elliott model. Numerical experiments demonstrate superior performance of the proposed RALC against the Consensus-Based Auction Algorithm (CBAA), the Probability- Tuned Market-based Allocation (PTMA), and the Repeated G- Prim auction (RGPrim) in communication degraded scenarios.