Department of Electronics

Advanced Signal Processing & Instrumentation Research Lab (ASPIRE)

Ambient sound is an important aspect of underwater acoustics and is an extremely efficient way to propagate energy through the ocean. Knowledge about the characteristics of ocean noise and its distribution relative to the location and movements of its sources is important for understanding the potential impacts of anthropogenic sound. Ambient noise in the ocean is the sound field against which signals must be detected. In the ambient noise field, it is often difficult to ascribe a particular sound to a localized source. The character of ambient noise results both from the characteristics of the multitude of contributing sound sources and the ability of sound to propagate efficiently from one location to another.

Noise in the ocean is the result of both natural and anthropogenic sources. Natural sources of noise include processes such as earthquakes, wind-driven waves, rainfall, bio-acoustic sound generation, and thermal agitation of the seawater. Anthropogenic noise is generated by a variety of activities, including commercial shipping; oil and gas exploration, development, and production (e.g. air-guns, ships, oil drilling); naval operations (e.g. military sonars, communications, and explosions); fishing (e.g. commercial/civilian sonars, acoustic deterrent, and harassment devices); research (e.g. air-guns, sonars, telemetry, communication, and navigation); and other activities such as construction, icebreaking, and recreational boating. Sources of anthropogenic noise are becoming both more pervasive and more powerful, increasing oceanic background noise levels as well as peak sound intensity levels.

Signal processing based on passive acoustic systems is exceedingly complex in the realms of the ocean due to several spectrally overlapping sources, the sheer dimensionality of the measurements as well as spatial and temporal fluctuations in the channel characteristics between successive measurements. Geometrical spreading loss, dispersion, scattering due to the presence of other underwater structures, channel inhomogeneities, stratification and environmental variability often mask the signal beyond a recognizable level. In such scenarios, the requirement often aims to detect and identify the biological sources from the ubiquitous clutter. High spatial, spectral and temporal resolution becomes essential in order to effectively identify biological vocalizations from the ambient acoustic field.

The complex acoustic environment of ocean characterized by a multitude of noise like signals and signal like noises, results in a bewildering mix of signal and noise, which demands the need of specialized systems with sophisticated signal processing techniques for extracting the signals of interest. These involve the use of beamforming techniques like Minimum variance distortionless response (MVDR) as well as Linear constraint minimum variance (LCMV) with direction of arrival (DOA) estimation using methods like Multiple Signal Classification (MUSIC), Minimum Norm-Method (MNM) and Estimation of Signal Parameters by Rotational Invariance Techniques (ESPRIT). The acoustical beamforming techniques for array signal processing produces algorithms, which involve massive Fast Fourier Transforms (FFT) that increases the levels of computational complexity.