Machine Source Localization of Tursiops truncatus Whistle-like Sounds in a Reverberant Aquatic Environment

Most research into bottlenose dolphins (Tursiops truncatus) capacity for communication has centered on tonal calls termed whistles, in particular individually distinctive contact calls referred to as signature whistles. While "non-signature" whistles exist, and may be important components of bottlenose dolphins communicative repertoire, they have not been studied extensively. This is in part due to the difficulty of attributing whistles to specific individuals, a challenge that has limited the study of not only non-signature whistles but the study of general acoustic exchanges among socializing dolphins. In this paper, we propose the first machine-learning-based approach to identifying the source locations of semi-stationary, tonal, whistle-like sounds in a highly reverberant space, specifically a half-cylindrical dolphin pool. We deliver estimated time-difference-of-arrivals (TDOAs) and normalized cross-correlation values computed from pairs of hydrophone signals to a random forest model for high-feature-volume classification and feature selection, and subsequently deliver the selected features into linear discriminant analysis, linear and quadratic Support Vector Machine (SVM), and Gaussian process models. In our 14-source-location setup, we achieve perfect accuracy in localization by classification and high accuracy in localization by regression (median absolute deviation of 0.66 m, interquartile range of 0.34 m - 1.57 m), with fewer than 10,000 features. By building a parsimonious (minimum-feature) classification tree model for the same task, we show that a minimally sufficient feature set is consistent with the information valued by a strictly geometric, time-difference-of-arrival-based approach to sound source localization. Ultimately, our regression models yielded better accuracy than the established Steered-Response Power (SRP) method when all training data were used, and comparable accuracy along the pool surface when deprived of training data at testing sites; our methods additionally boast improved computation time and the potential for superior localization accuracy in all dimensions with more training data.