Multi-target 2D tracking method for singing humpback whales using vector sensors

Area of Science:

• Marine bioacoustics
• Array signal processing
• Animal acoustic tracking

Background:

• Acoustic vector sensors measure acoustic pressure and particle motion for wave direction estimation.
• Triangulation with vector sensors can locate acoustic sources, but tracking multiple simultaneous sources is challenging.
• Identifying and linking azimuthal measurements from multiple sensors to individual sources requires advanced methods.

Purpose of the Study:

• To demonstrate the use of 2D vector sensors for tracking multiple, simultaneously singing humpback whales.
• To develop a method for linking azimuthal tracks between sensors for accurate source identification.
• To localize and estimate swimming speeds of individual whales using acoustic data.

Main Methods:

• Deployment of 2D vector sensors off the coast of western Maui.
• Generation of azimuthal tracks from individual whale vocalizations using time-frequency representations (azigrams).
• Incorporation of acoustic transport velocity estimates and cross-correlation of azigram features for track linking.
• Localization and tracking in latitude and longitude using data from a minimum of two vector sensors.

Main Results:

• Successful generation of high-quality azimuthal tracks for simultaneously singing humpback whales.
• Effective linking of azimuthal tracks between sensors through cross-correlation of azigram features.
• Accurate two-dimensional tracks and positional uncertainties for six singing whales.
• Estimation of swimming speeds derived from high-quality tracks.

Conclusions:

• 2D vector sensors are effective tools for tracking multiple simultaneous acoustic sources like singing whales.
• Cross-correlation of azigrams with acoustic transport velocity estimates enables reliable track association between sensors.
• This methodology allows for precise localization and behavioral analysis of marine mammals.