A bio-inspired distributed approach for searching underwater acoustic source using a team of AUVs

We present a bio-inspired distributed algorithm which is a fusion of two distinct animal behaviours to solve three different underwater search missions. One of the two constituent control modules is called target-drive which models the hypothesized behaviour of a fish-larva searching for a coral reef using acoustic cues. Target-drive helps a single Autonomous Underwater Vehicle (AUV) to adjust its heading towards the acoustic source. The other control module called group-cohesion, mimics movements of a golden shiner (Notemigonus crysoleucas) in a school of fish. The proposed approach only relies on implicit communication and achieves target convergence only by assuming a single on-board hydrophone and location estimate of AUV's neighbours. The effects of varying key parameters such as group-size and neighbourhood-radius on convergence times have been thoroughly investigated. We present a preliminary analysis of the algorithm's performance which shows promise in solving problems employing small teams of AUVs in terms of convergence times and non-existent inter-agent communication.