Probabilistic spatially-divided multiple access in underwater acoustic sparse networks

Deploying Autonomous Underwater Vehicles (AUVs) is a necessity to enable a range of civilian/military underwater applications; yet, achieving a reliable coordination among the vehicles is a challenging issue due to the time- and space-varying characteristics of the acoustic communication channel. The design of a Medium Access Control (MAC) based on a probabilistic Space Division Multiple Access (SDMA) method for short/medium distances (less than 2 km 2 km ) is presented. This method considers the inherent vehicle position uncertainty due to the inaccuracies in models and the drift of the vehicles. It minimizes the acoustic interference statistically by considering the angular position of neighboring vehicles via a two-step estimation and by keeping the transmitter antenna's beamwidth of each vehicle at an optimal value. Such value is chosen considering three contrasting goals, i.e.: (i) spreading the signal beam towards the vehicle to combat position uncertainty using a coarse estimation; (ii) focusing the beam to reduce acoustic energy dispersion through a fine estimation; and (iii) minimizing interference to other vehicles. Simulation results in a sparse underwater network show that this approach mitigates interference, reduces the probability of retransmission, and achieves higher data rates over conventional underwater MAC techniques.