Single receiver target localization in mobile marine acoustic telemetry

Many marine ecological studies track animal movement and migration using acoustic telemetry, in which animals are implanted with acoustic transmitters broadcasting a unique ID code. Receivers detect and decode these transmissions and provide an estimate of the animal’s location. As transmissions can be detected at distances of hundreds of meters to even several kilometers, localization performed using binary detection information is very coarse. Fine-scale localization can be performed using time difference of arrival systems, but such systems are costlier and require a greater number of receivers. We present a Bayesian method for target localization using only binary detection information collected by a single receiver mounted on a mobile platform. The approach reduces error in the location estimate by considering the distance-dependent probability of detection (detection efficiency curve), which must first be characterized. Stationary receivers are used to infer failed detections. This localization method is then used to inform waypoint selection for the mobile platform, using the second nondominated sorting genetic algorithm considering the tradeoff between the travel distance of the platform and localization error. Simulation results examine the effects of detection efficiency curve characteristics, number of transmissions, and travel distance on the localization error. The proposed method is further supported by experimental results from field trials conducted with an autonomous surface vehicle in Lake Huron.