Feature ArticleEnvironmentally Neutral Biomimetic Waveforms for ASW Sonar
By Peter Dobbins
Ultra Electronics Sonar Systems
There is a rapidly expanding requirement in the defense, offshore, renewable energy, research and leisure sectors to reduce the impact of man-made sound, including active sonar transmissions, on marine mammals. This is driven partly by public interest in these animals, but mainly by legislation such as the U.S. Marine Mammal Protection Act, the U.K. Offshore Marine Conservation Regulations, and similar regulatory and licensing requirements throughout the world.
Typically, such requirements are met using visual monitoring by marine mammal observers or passive acoustic monitoring based on listening for vocalizations such as the echolocation clicks of dolphins and porpoises or the low-frequency calls of baleen whales. If animals are detected within a specified range, some form of mitigating action such as shutting down the sound source is then necessary.
However, some marine species do not vocalize, and in general, it may not be possible to ensure the absence of marine mammals before commencing transmission. Therefore it is desirable to look for forms of sonar transmission that are potentially less harmful to marine life without needing to reduce the transmission level or shut down completely.
One way this might be achieved is to use signal waveforms derived from naturally occurring sounds, such as the vocalizations of the animals themselves: biomimetic waveforms. It might be expected that such sounds would appear less threatening (or at least more familiar), thus reducing possible abnormal behavioral impacts.
The interaction of marine mammals and man-made or anthropogenic sound is a subject of some contention. The understanding of detailed interactions is relatively poor and largely restricted to a few species. However, there are instances where man-made sounds have been demonstrated to have adverse effects on marine mammals, and this includes sonar transmissions.
There are widely accepted guidelines relating to exposure criteria for injury in the form of temporary or permanent threshold shift. However, behavioral impacts are difficult to measure or predict. The challenge is to distinguish a significant behavioral response from an insignificant momentary alteration in behavior. It is reasonable to assume that the biomimetic waveforms under consideration are likely to minimize behavioral reactions, but this assumption would need to be verified before any such waveforms could be used in service.
For physical impacts, given a specified waveform, transmission level and other sonar parameters, the sound pressure level in the water around an active sonar can be computed. Comparison between the sound pressure levels and the exposure criteria can then be used to assess the likelihood of physical impact, either temporary or permanent threshold shift, for a variety of mammals with different hearing ranges. This assessment can be based on the instantaneous sound pressure level or, possibly more realistically, the cumulative exposure of an animal to the sound field for an extended period as it moves around, referred to as the sound exposure level (SEL).
In order to weigh the performance of conventional sonar signals against the biomimetic waveforms, a simple metric may be applied: The potential detection performance can be compared for signals adjusted to obtain the same SEL for a given marine mammal target in a specified environment. This equates to both waveforms having the same ambient noise limited detection performance.
Marine mammals produce a variety of vocalizations, but this article focuses on the echolocation clicks produced by dolphins and other odontocetes (toothed whales), mainly because these are active sonar signals, whereas many of the vocalizations associated with these animals are for communications. However, this does not preclude the possibility of using biomimetic communication signals for sonar.
The original inspiration for these novel signals came from the analysis of clicks from bottlenose dolphins (Tursiops truncatus). The pulses are of very short duration, between 50 and 80 microseconds, and spectrograms computed using short-time fractional Fourier transforms clearly show that the signal comprises two short downward chirps. However, although the double chirp structure seems typical of bottlenose waveforms, few other species have been studied, except for sperm whale (Physeter macrocephalus) clicks in which, once again, the double chirp structure is evident, although much lower in frequency and extended in duration.
Although not proven, it seems that the double chirp structure may be in widespread use for odontocete echolocation waveforms and scalable in both time and frequency. This signal structure might appear less threatening to most mammals and thus have a lower behavioral impact. Therefore, a biomimetic signal model was implemented based on two linear-frequency-modulated chirps. In this implementation, a waveform is fully defined by the frequency range and duration of the two chirps, along with the delay between the first and second. A waveform representative of sperm whale clicks has been chosen for the sonar performance analysis presented in this article.
For comparison with the synthesized echolocation click, a typical anti-submarine warfare (ASW), mid-frequency sonar waveform operating at a center frequency of seven kilohertz with a 500-millisecond, 100-hertz bandwidth chirp pulse was used as an example. For this simple illustration, the same typical sonar source level is used for both conventional and biomimetic waveforms, the sonar projector is assumed omnidirectional, and the pulse repetition frequency is one pulse per 10 seconds.
The instantaneous SEL can be calculated for a single pulse by integrating the square of the pressure waveform over the duration of the pulse, having determined the sound pressure level received by a mammal at a specified position relative to the sonar. The cumulative SEL can then be obtained from these values for a mammal swimming over a specified path simply by summing the resultant SELs for each ping transmitted for the duration of interest. To continue this article please click here.
Peter Dobbins is principal scientist with the imaging sonar group at Ultra Electronics. He received his Ph.D. in underwater acoustics from the University of Bath. His fields of activity include bioacoustics, using acoustics to study marine mammals and applying biologically inspired techniques to man-made sonar.