Automated extraction of baleen whale calls based on the pseudo-Wigner-Ville distribution.

Baleen whales produce a wide variety of frequency-modulated calls. Extraction of the time-frequency (TF) structures of these calls forms the basis for many applications, including abundance estimation and species recognition. Typical methods to extract the contours of whale calls from a spectrogram are based on the short-time Fourier transform and are, thus, restricted by a fixed TF resolution. Considering the low-frequency nature of baleen whale calls, this work represents the contours using a pseudo-Wigner-Ville distribution for a higher TF resolution at the cost of introducing cross terms. An adaptive threshold is proposed followed by a modified Gaussian mixture probability hypothesis density filter to extract the contours. Finally, the artificial contours, which are caused by the cross terms, can be removed in post-processing. Simulations were conducted to explore how the signal-to-noise ratio influences the performance of the proposed method. Then, in experiments based on real data, the contours of the calls of three kinds of baleen whales were extracted in a highly accurate manner (with mean deviations of 5.4 and 0.051 Hz from the ground-truth contours at sampling rates of 4000 and 100 Hz, respectively) with a recall of 75% and a precision of 78.5%.

Robust unsupervised Tursiops aduncus whistle enhancement based on complete ensembled empirical optimal envelope local mean decomposition with adaptive noise.

Whistle enhancement is an essential preprocessing step in studying dolphin behavior and population distributions. We propose a robust unsupervised whistle enhancement scheme based on improved local mean decomposition using adaptive noise estimation and logarithmic spectral amplitude. First, to further mitigate the mode aliasing problem effect in whistle signal decomposition and achieve better spectral separation of modes, we present a complete ensembled empirical optimal envelope local mean decomposition with adaptive noise algorithm. According to the envelope characteristics of the whistle signals, the proposed algorithm optimally and adaptively decomposes the noisy signal into product functions (PFs) with amplitude and frequency modulation. Second, the whistle enhancement framework consists of the improved minima-controlled recursive averaging for adaptive noise estimation, optimally modified log-spectral amplitude for each noisy product function enhancement, and the Hurst index for reconstructing pure whistle signal estimations with the least damaged PFs. Finally, the proposed scheme is applied to a dataset of long calls from two Tursiops aduncus individuals. After constructing the pure whistle dataset, the experimental results show that the proposed scheme performs better than other compared whistle enhancement schemes under different signal-to-noise ratios.

Effects of receiving position and shell material on in-band full-duplex underwater communications' self-interference signal.

Despite the recent intensive research on adaptive algorithms for self-interference (SI) cancellation (SIC) in in-band full-duplex (IBFD) underwater acoustic communication (UWAC), there has been relatively little exploration of how the IBFD-UWAC modem shell affects the SI signal. This paper analyzes the effects of the shell material and the near-end receiver position on the SI signal. The analysis is done with a two-dimensional finite-element model in a free-field simulation environment, which combines the differential equation of motion and the time-dependent solver. The SI signal strength around the modem shell in the far-field conditions is obtained. The simulation and pool experiment results both show that (i) the strength of the received SI signal is lowest when the near-end receiver is on a line extending from the shell's geometric center perpendicularly to its central axis and (ii) a shell material with a high elastic coefficient is more conducive to suppressing the SI signal. A pool experiment showed that changing the spatial position of the near-end receiver and the shell material from aluminum to stainless steel enhanced the SIC performance of the IBFD-UWAC system by at least 12 and 4 dB, respectively.

Robust unsupervised Tursiops aduncus whistle-event detection using gammatone multi-channel Savitzky-Golay based whistle enhancement.

Detecting whistle events is essential when studying the population density and behavior of cetaceans. After eight months of passive acoustic monitoring in Xiamen, we obtained long calls from two Tursiops aduncus individuals. In this paper, we propose an algorithm with an unbiased gammatone multi-channel Savitzky-Golay for smoothing dynamic continuous background noise and interference from long click trains. The algorithm uses the method of least squares to perform a local polynomial regression on the time-frequency representation of multi-frequency resolution call measurements, which can effectively retain the whistle profiles while filtering out noise and interference. We prove that it is better at separating out whistles and has lower computational complexity than other smoothing methods. In order to further extract whistle features in enhanced spectrograms, we also propose a set of multi-scale and multi-directional moving filter banks for various whistle durations and contour shapes. The final binary adaptive decisions at frame level for whistle events are obtained from the histograms of multi-scale and multi-directional spectrograms. Finally, we explore the entire data set and find that the proposed scheme achieves the highest frame-level F1-scores when detecting T. aduncus whistles than the baseline schemes, with an improvement of more than 6%.

Automated classification of tursiops aduncus whistles based on a depth-wise separable convolutional neural network and data augmentation.

Whistle classification plays an essential role in studying the habitat and social behaviours of cetaceans. We obtained six categories of sweep whistles of two Tursiops aduncus individual signals using the passive acoustic mornitoring technique over a period of eight months in the Xiamen area. First, we propose a depthwise separable convolutional neural network for whistle classification. The proposed model adopts the depthwise convolution combined with the followed point-by-point convolution instead of the conventional convolution. As a result, it brings a better classification performance in sample sets with relatively independent features between different channels. Meanwhile, it leads to less computational complexity and fewer model parameters. Second, in order to solve the problem of an imbalance in the number of samples under each whistle category, we propose a random series method with five audio augmentation algorithms. The generalization ability of the trained model was improved by using an opening probability for each algorithm and the random selection of each augmentation factor within specific ranges. Finally, we explore the effect of the proposed augmentation method on the performance of our proposed architecture and find that it enhances the accuracy up to 98.53% for the classification of Tursiops aduncus whistles.

Symmetric Connectivity of Underwater Acoustic Sensor Networks Based on Multi-Modal Directional Transducer.

Topology control is one of the most essential technologies in wireless sensor networks (WSNs); it constructs networks with certain characteristics through the usage of some approaches, such as power control and channel assignment, thereby reducing the inter-nodes interference and the energy consumption of the network. It is closely related to the efficiency of upper layer protocols, especially MAC and routing protocols, which are the same as underwater acoustic sensor networks (UASNs). Directional antenna technology (directional transducer in UASNs) has great advantages in minimizing interference and conserving energy by restraining the beamforming range. It enables nodes to communicate with only intended neighbors; nevertheless, additional problems emerge, such as how to guarantee the connectivity of the network. This paper focuses on the connectivity problem of UASNs equipped with tri-modal directional transducers, where the orientation of a transducer is stabilized after the network is set up. To efficiently minimize the total network energy consumption under constraint of connectivity, the problem is formulated to a minimum network cost transducer orientation (MNCTO) problem and is provided a reduction from the Hamiltonian path problem in hexagonal grid graphs (HPHGG), which is proved to be NP-complete. Furthermore, a heuristic greedy algorithm is proposed for MNCTO. The simulation evaluation results in a contrast with its omni-mode peer, showing that the proposed algorithm greatly reduces the network energy consumption by up to nearly half on the premise of satisfying connectivity.

Estimation of Overspread Underwater Acoustic Channel Based on Low-Rank Matrix Recovery.

In this paper, the estimation of overspread, i.e., doubly spread underwater acoustic (UWA) channels of strong dispersion is considered. We show that although the UWA channel dispersion causes the degeneration of channel sparsity, it leads to a low-rank structure especially when the channel delay-Doppler-spread function is separable in delay and Doppler domain. Therefore, we introduce the low-rank criterion to estimate the UWA channels, which can help to improve the estimation performance in the case of strong dispersion. The estimator is based on the discrete delay-Doppler-spread function representation of channel, and is formulated as a low-rank matrix recovery problem which can be solved by the singular value projection technique. Simulation examples are carried out to demonstrate the effectiveness of the proposed low-rank-based channel estimator.

Three-dimensional finite element simulation of acoustic propagation in spiral bubble net of humpback whale.

In 2004, Leighton hypothesized that the acoustic calls emitted by humpback whales when feeding using bubble nets, may enhance the effectiveness of the net in confining prey (such as herring) by forming a "wall of sound" with a quiet zone within. Modelling of the acoustics of this phenomenon was previously restricted to 2D; this paper conducts a 3D model of the propagation of signals resembling those emitted by humpback whales when bubble netting, projected into an upward spiral bubble net which data to date suggest is the accurate form for the bubble net in 3D space. In this study, the feeding calls were analyzed in the time-frequency domain to extract acoustic information sufficient to allow modeling of the resulting spatial distribution of acoustic pressure and particle velocity, and how they vary over the duration of the call. Sound propagation in the bubble net was described by using a linear steady-state formulation for an effective medium of bubbly water. Using the predicted attenuation, phase velocity and density in bubbly water, a 3D finite element model was constructed to numerically simulate the upward-spiral bubble net which consists of a mixture of bubbles that exhibit a range of radii. The acoustic pressure field and particle motion field were both calculated within the bubble net. The simulation results show that the energy of the whale feeding call could be effectively focused in the bubble net, generating intensive sound pressure and particle motion fields in the bubbly arm of the net, but with some "quiet" regions closer to the center of the net, as Leighton hypothesized. Furthermore, when the hearing ability of herring is taken into consideration, the results suggest that this acoustic focusing effect could be a plausible factor in trapping them in the bubble net. It also allows speculation on the possible enhancements that the time-varying nature of the call during feeding could give to the whale in this mechanism for the bubble net feeding by humpback whales.

Doppler scale estimation for varied speed mobile frequency-hopped binary frequency-shift keying underwater acoustic communication.

An optimized Doppler scale factor estimation method based on wavelet de-noising is proposed in this paper to improve the performance of frequency-hopped binary frequency-shift keying mobile underwater acoustic communication. The method is with regard to the Doppler scale factors of multiple symbols as a whole and focuses on the continuous variation overall trends, rather than the association between two adjacent symbols. The proposed method was demonstrated in a shallow water experiment conducted in November 2018 at Songhua River in Heilongjiang, China. Experimental data collected showed a bit error rate less than 10-4 with a data rate of 65.41 bps at a moving speed of 2.0-6.2 m/s over the horizontal range of 0.68-1.6 km.

Digital Self-Interference Cancellation for Asynchronous In-Band Full-Duplex Underwater Acoustic Communication.

To improve the throughput of underwater acoustic (UWA) networking, the In-band full-duplex (IBFD) communication is one of the most vital pieces of research. The major drawback of IBFD-UWA communication is Self-Interference (SI). This paper presents a digital SI cancellation algorithm for asynchronous IBFD-UWA communication system. We focus on two issues: one is asynchronous communication dissimilar to IBFD radio communication, the other is nonlinear distortion caused by power amplifier (PA). First, we discuss asynchronous IBFD-UWA signal model with the nonlinear distortion of PA. Then, we design a scheme for asynchronous IBFD-UWA communication utilizing the non-overlapping region between SI and intended signal to estimate the nonlinear SI channel. To cancel the nonlinear distortion caused by PA, we propose an Over-Parameterization based Recursive Least Squares (RLS) algorithm (OPRLS) to estimate the nonlinear SI channel. Furthermore, we present the OPRLS with a sparse constraint to estimate the SI channel, which reduces the requirement of the length of the non-overlapping region. Finally, we verify our concept through simulation and the pool experiment. Results demonstrate that the proposed digital SI cancellation scheme can cancel SI efficiently.

Dolphin Sounds-Inspired Covert Underwater Acoustic Communication and Micro-Modem.

A novel portable underwater acoustic modem is proposed in this paper for covert communication between divers or underwater unmanned vehicles (UUVs) and divers at a short distance. For the first time, real dolphin calls are used in the modem to realize biologically inspired Covert Underwater Acoustic Communication (CUAC). A variety of dolphin whistles and clicks stored in an SD card inside the modem helps to realize different biomimetic CUAC algorithms based on the specified covert scenario. In this paper, the information is conveyed during the time interval between dolphin clicks. TMS320C6748 and TLV320AIC3106 are the core processors used in our unique modem for fast digital processing and interconnection with other terminals or sensors. Simulation results show that the bit error rate (BER) of the CUAC algorithm is less than 10 - 5 when the signal to noise ratio is over ‒5 dB. The modem was tested in an underwater pool, and a data rate of 27.1 bits per second at a distance of 10 m was achieved.

Elastic feature of cylindrical shells extraction in time-frequency domain using biomimetic dolphin click.

A dolphin's biosonar may effectively discriminate subtle differences among targets. In order to investigate the possible physical mechanism of target discrimination, in this study, a finite element model excited by a biomimetic click pulse was proposed. The acoustic scattering field and stress distribution of a stainless steel shell were simulated. The biomimetic click experiments were then conducted to verify the theoretical predictions in an anechoic tank. The experimental results showed a good agreement with the model simulations. Furthermore, the elastic time-frequency features of three cylindrical shells with different wall thickness were obtained using a fractional Fourier transform filter to eliminate specular reflection and cross-term interference. To compare discrimination capacity of the time-frequency features with and without the specular reflection, a time-frequency correlator was applied to calculate the correlation coefficient between different shells. The results indicated that the time-frequency features can be represented in high resolution with less cross-term interference, and these features without specular reflection showed a good capacity to discriminate the shells with different wall thickness.

Covert underwater acoustic communication using dolphin sounds.

In November 2012, an experiment demonstrating biological mimicry method for covert underwater acoustic communication (UAC) was conducted at Lianhua Lake in Heilongjiang China. Dolphin whistles were used for synchronization while dolphin clicks were used as information carrier. The time interval between dolphin clicks conveys the information bits. Channel estimates were obtained with matching pursuit (MP) algorithm, which is useful for sparse channel estimation. Adaptive RAKE Equalization was employed at the receiver. Bit error rates were less than 10(-4) with 37 bits per second data rate in the lake trial.