An adaptive MAC-agnostic ranging scheme for underwater acoustic mobile networks

Abstract

In the last ten years several simulation studies on Autonomous Underwater Vehicle (AUV) swarm fleet formation have been performed, and some preliminary sea demonstrations of proof-of-concept prototypes were carried out. However, their actual realization is hindered by the challenges imposed by the underwater acoustic channel and the difficulties of keeping track of the vehicles’ positions due to the long latency required by traditional Two-Way Travel-Time (TWTT) ranging measurements, that require a specific signaling, hence limiting the throughput of the underwater network. Although One-Way Travel-Time (OWTT) halves the latency, it requires a high precision oscillator, such as an atomic clock or an oven controlled crystal oscillator (OCXO), to be installed in each modem processing unit: while atomic clocks are still very expensive, OCXO are very power demanding, making their application to underwater acoustic networks not always possible, especially in the case of low cost vehicle swarms.In this paper we present a network protocol stack able to perform ranging and localization within the communication task in underwater acoustic networks, limiting the network overhead. Specifically, new layers have been added to the preexisting DESERT Underwater protocol stack to perform the ranging tasks without compromising the correct operation of the communication network. A ranging layer is placed on top of the Medium Access Control (MAC) scheme, allowing the latter to be changed according to the network topology and requirements. This MAC-agnostic ranging layer is further optimized by adapting the amount of data transmitted according to the channel state, and the ranging entries inserted in the data packets according to the least recent information transmitted, hence minimizing the Age of Information. Simulation results obtained with the DESERT Underwater Framework show how this layer allows all AUVs in the swarm to know their distance from every other node in the network, thus limiting the probability of vehicle collisions and allowing better mission coordination.