Empirical Evaluation of Auction-Based Coordination of AUVs in a Realistic Simulated Mine Countermeasure Task

Abstract

In this work, we evaluate performance of our distributed cooperation framework, DEMiR-CF, for Naval Mine Countermeasure missions on the US NAVY’s ALWSE-MC simulator against different contingencies that may arise run time. Our cooperation framework integrates a distributed task allocation scheme, coordination mechanisms and precaution routines for multirobot team execution. Its performance has been demonstrated in Multi-robot Multi-target exploration and Object Construction domains. Marine applications provide additional challenges such as noisy communication, position uncertainty and the likelihood of robot failures. There is a high probability that the initial assignments are subject to change during run time, in these kinds of environments. Our framework ensures robust execution and efficient completion of missions against several different types of failures. Preliminary results for MCM missions are promising in the sense of mission completion, and AUV paths are close to optimal in the presence of uncertainties.