Frequency domain tracking of passive vessel harmonics.

Abstract

This paper presents a method for passive acoustic detection and tracking of small vessels in noisy, shallow water marine environments. Passive spectra of boats include broadband noise as well as tones that are harmonics of the engine speed and shaft/propeller rotation. Past work suggests that the location in frequency and the relative amplitudes of these harmonics can be used to determine specific characteristics of the vessel such as the number of blades on the propeller and engine type/speed. However, the low signal to noise ratio of quiet targets and Doppler shifts incurred because of source and receiver motion complicate the identification of these tones in the lofargram. To address this issue, a combined detection and tracking approach is proposed in which intermittent and wandering harmonic content is tracked with a multi-dimensional Kalman filter. Results from recorded passive signatures from several classes of vessels in marine and freshwater environments in the Pacific Northwest are presented and discussed.

Extraction of small boat harmonic signatures from passive sonar

“…The value W is a measure of how well the data (5) fits the signal replica model (7). A value of W equal to one would mean that the data perfectly matches the model.…”

“…Passive spectra of boats include broadband noise as well as tonals due to the harmonics of the engine speed and shaft/ propeller rotation. 7 The algorithm developed here extracts the harmonic features to facilitate the exploration of the relationship between these features and the identification of specific boats. These features consist of harmonic amplitudes, SNRs, and the fundamental frequencies of the boat noise.…”

A fast algorithm for maximum likelihood-based fundamental frequency estimation

“…Such periodic signals are encountered in a wide range of applications such as music and speech processing [1,2], sonar [3,4], and electrocardiography [5]. Although the signals in these applications are all real-valued, these are often transformed into down-sampled analytic signals [6] for two reasons.…”

Accurate Estimation of Low Fundamental Frequencies From Real-Valued Measurements

“…Also in the analysis of some bird calls and various other biological signals, like vital signs [1], such signals can be encountered. Moreover, they occur in radar applications for rotating targets [2] and in passive detection, localization, and identification of boats and helicopters [3]. It is then also not surprising that a host of methods have been proposed over the years including methods based on the principles of maximum likelihood, least-squares (LS), and weighted least-squares (WLS) [4]- [8], auto-/crosscorrelation and related methods [9]- [13], linear prediction [14], filtering [2], [15]- [17], and subspace methods [18], [19].…”