Sperm whales communicate in short sequences of clicks known as codas. We present WhAM (Whale Acoustics Model), the first transformer-based model capable of generating synthetic sperm whale codas from any audio prompt. WhAM is built by finetuning VampNet, a masked acoustic token model pretrained on musical audio, using 10k coda recordings collected over the past two decades. Through iterative masked token prediction, WhAM generates high-fidelity synthetic codas that preserve key acoustic features of the source recordings. We evaluate WhAM's synthetic codas using Fréchet Audio Distance and through perceptual studies with expert marine biologists. On downstream tasks including rhythm, social unit, and vowel classification, WhAM's learned representations achieve strong performance, despite being trained for generation rather than classification. Our code is available at https://github.com/Project-CETI/wham

To protect underwater acoustic communications from interception, the exchange of encryption keys is necessary. Since underwater devices can be compromised, generating keys on site is a better option than predefined keys. In this paper, we present a solution that utilizes the characteristics of the underwater acoustic channel impulse response (CIR), which is highly variable in both space and time, while ensuring consistency between the communicating nodes (Alice and Bob). To compensate for temporal variations, the key is calculated from the CIR feature’s distribution parameters, while the hard key is determined using a K-means strategy. To achieve key agreement, we exploit the long propagation delay in the underwater CIR and let Alice and Bob transmit simultaneously while their packets fly past each other. Due to the channel’s reciprocity, this simultaneous transmission ensures that the same CIR is estimated at both ends of the communication link. Simulation and sea experiment results show that it is possible to extract at least three times as many secret bits as we would with a uniform quantizer. The results show a high matching rate between Alice and Bob and a high Hamming distance to Eve’s key. For reproducibility, we share the CIRs from the sea trials.

We describe a biomimicking interception scheme tailored to Underwater Acoustic Communications (UAC), which aims at separating authentic and biomimicking signals. Our interceptor leverages the expected stability of the biomimicking sources, as opposed to vocalizations by marine fauna, which are expected to move fast and rapidly change orientation. Consequently, the channel impulse response (CIR) of the link between a receiver and a biomimicking source is expected to be much more stable than those corresponding to actual vocalizations. We quantify this stability by testing the randomness of the representation of the CIRs. The latter are represented by two similarity metrics: the cross-correlation and the sample entropy between adjacent CIRs features. We offered two interception measures: 1) testing the similarity between a Gaussian distribution and the distribution of the similarity measures using the Kullback–Leibler divergence (KLD) criteria for quantification, and 2) the minimum number of clusters to effectively segment the similarity measures as a point cloud. Results from simulations for artificial signals mimicking dolphin whistles and the outcomes of a lake trial demonstrate the effectiveness of our biomimicking interceptor.

Sperm whales (Physeter macrocephalus) navigate underwater with a series of impulsive, click-like sounds known as echolocation clicks. These clicks are characterized by a multipulse structure (MPS) that serves as a distinctive pattern. In this work, we use the stability of the MPS as a detection metric for recognizing and classifying the presence of clicks in noisy environments. To distinguish between noise transients and to handle simultaneous emissions from multiple sperm whales, our approach clusters a time series of MPS measures while removing potential clicks that do not fulfil the limits of inter-click interval, duration and spectrum. As a result, our approach can handle high noise transients and low signal-to-noise ratio. The performance of our detection approach is examined using three datasets: seven months of recordings from the Mediterranean Sea containing manually verified ambient noise; several days of manually labelled data collected from the Dominica Island containing approximately 40,000 clicks from multiple sperm whales; and a dataset from the Bahamas containing 1,203 labelled clicks from a single sperm whale. Comparing with the results of two benchmark detectors, a better trade-off between precision and recall is observed as well as a significant reduction in false detection rates, especially in noisy environments. R1C1a To ensure reproducibility, we provide our database of labelled clicks along with our implementation code.

This paper proposes a simultaneous localization and mapping (SLAM) solution that combines inputs from sonar and optical images. Our solution for this navigation problem is solved by matching objects detected by the AUV's camera with objects identified within the SAS image. In particular, upon detecting an object using the AUV's camera, e.g., an underwater structure or even a rock, we rank the similarity of this object to all objects found within the SAS image. Taking a SLAM approach, the decision for the best match, and thus the position of the AUV within the SAS image, is made based not only on the current detected object, but also on similarity ranking previously detected objects while accounting for the AUV's self-measured heading. The solution is handled by a tracking mechanism that considers the possible positions of the objects within the SAS image as problem states, and modeling the state relations using a hidden Markov chain. Experimental results on real sonar and optical images implemented within a simulated environment show high accuracy in matching the location of an AUV within an SAS image in multiple scenarios.

The calibration of sensors stationed along a cable in marine observatories is a time-consuming and expensive operation that involves taking the mooring out of the water periodically. In this paper, we present a method that allows an underwater vehicle to approach a mooring, in order to take reference measurements along the cable for in-situ sensor calibration. We use the vehicle's Mechanically Scanned Imaging Sonar (MSIS) to identify the cable's reflection within the sonar image. After pre-processing the image to remove noise, enhance contour lines, and perform smoothing, we employ three detection steps: 1) selection of regions of interest that fit the cable's reflection pattern, 2) template matching, and 3) a track-before-detect scheme that utilized the vehicle's motion. The later involves building a lattice of template matching responses for a sequence of sonar images, and using the Viterbi algorithm to find the most probable sequence of cable locations that fits the maximum speed assumed for the surveying vessel. Performance is explored in pool and sea trials, and involves an MSIS onboard an underwater vehicle scanning its surrounding to identify a steel-core cable. The results show a sub-meter accuracy in the multi-reverberant pool environment and in the sea trial. For reproducibility, we share our implementation code.

Sperm whales are characterized by the inter-pulse interval (IPI) of their vocalization, which is a function of the whales’ size. In this paper, we propose a new methodology for detecting and extracting features from the IPI of sperm whale clicks. Our IPI estimator is based on the slope of the signal’s phase spectrum, which takes on zero values whenever a pulse is observed. To manage outliers, we propose a segmentation algorithm for the choice of the most consistent IPI subset within a time buffer. Finally, a linear fit and standard deviation are calculated over the segmented subset. The proposed approach is demonstrated over data of recorded ambient noise (false alarm) and real sperm whale signals obtained in two designated sea experiments. Performance is compared with two widely used IPI-estimation benchmarks and with a waveform-based approach representative. Our results show that the proposed approach yields features with the highest separation.

With the growing use of underwater acoustic communications and the recent adoption of standards in this field, it is becoming increasingly important to secure messages against eavesdroppers. In this paper, we focus on a physical-layer security solution to generate sequences of random bits (keys) between two devices (Alice and Bob) belonging to an underwater acoustic network (UWAN); the key must remain secret to a passive eavesdropper (Eve) not belonging to the UWAN. Our method is based on measuring the propagation delay of the underwater acoustic channel over multiple hops of the UWAN: this harvests the randomness in the UWAN topology and turns the slow sound propagation in water into an advantage against eavesdropping. Our key generation protocol includes a route discovery handshake, whereby all UWAN devices at intermediate hops accumulate their message processing delays. This enables Alice and Bob to compute the actual propagation delays along each route and to map such information to a sequence of bits. Finally, from these bit sequences, Alice and Bob obtain a secret key. We analyze the performance of the protocol theoretically and assess it via extensive simulations and field experiments.

Long-range underwater acoustic communication (LR-UWAC) is an essential technique for applications such as the control of unmanned underwater vehicles, gliders, or tactical submarines for long-term monitoring tasks. While techniques for estimating the channel impulse response (CIR) of short-range UWAC have been successfully applied, this is not the case for LR-UWAC. Here, the main challenge is to handle the diverse channel structure, which can be of a sparse or a non-sparse structure. In this paper, we propose a robust channel estimator for LR-UWAC. Our solution includes a switching mechanism to classify the channel as either sparse or non-sparse. In its sparse form, the LR-UWAC channel is characterized by a long time delay spread, time-invariant characteristics, and a block structure. Thus, we propose a non-trivial combination of the block subspace pursuit and the distributed subspace pursuit algorithms to exploit the channel’s block structure. For non-sparse LR-UWAC channel structures, we approximate the received signal as Gaussian, and derive a maximum-likelihood (ML) estimator to separate the non-resolvable multipath. Extensive numerical simulation and results from two long-range sea experiments demonstrate the efficiency of our approach in identifying the channel’s structure, and to accurately estimate the channel compared to state-of-the-art benchmarks.

The accurate detection and quantification of submerged targets has been recognized as a key challenge in marine exploration, one that traditional census approaches cannot handle efficiently. Here we present a deep learning approach to detect the pattern of a moving fish from the reflections of an active acoustic emitter. To allow for real-time detection, we use a convolutional neural network, which provides the simultaneous labeling of a large buffer of signal samples. This allows to capture the structure of the reflecting signal from the moving target and to separate it from clutter reflections. We evaluate system performance both on synthetic (simulated) data, as well as on real data recorded over 50 sea experiments in a variety of sea conditions. When tested on real signals, the network trained on simulated patterns showed non-trivial detection capabilities, suggesting that transfer learning can be a viable approach in these scenarios, where tagged data is often lacking. However, training the network directly on the real reflections with data augmentation techniques allowed to reach a more favorable precision-recall trade-off, approaching an ideal detection bound. We also evaluate an alternative model based on recurrent neural networks which, despite exhibiting slightly inferior performance, could be applied in scenarios requiring on-line processing of the reflection sequence.

An effective method for detecting the presence of dolphins is by using passive acoustic monitoring (PAM), where pod size indications can be estimated by counting individual whistles. The detection of dolphin whistles is commonly applied on a time-frequency representation, followed by denoising and whistle tracking to evaluate the number of whistles. However, due to harmonics, multipath and time-varying signal-to-noise ratio, a single dolphin whistle may be associated with multiple whistle-traces. Thus, as a first step towards evaluating dolphins' abundance, our goal is to cluster individual whistle traces into unique whistles. Our scheme measures the similarity between each pair of whistle traces, and estimates the likelihood of whistle traces sharing the same cluster. Clustering is formalized as an optimization problem, aims to maximize the stability of clusters. Formalizing the problem as a minimal-cut optimization on a graph provides an effective solution based on spectral decomposition of the graph-Laplacian. Our model of the likelihood sharing cluster provides a physically-meaningful method to calculate the graph's connectivity parameters, thereby leading to a robust blind clustering. Based on numerical simulations and real recordings of dolphin whistles at sea, we demonstrate the applicability of our solution and its advance beyond alternative approaches.

A key technology in the movement tracking of marine animals is localization using acoustic transmitters. These are attached to marine animals and are detected by an array of receivers. Then, offline localization is performed by multilateration. However, due to the transmitter's low power and environmental conditions, emissions may be detected by only a limited number of receivers, causing localization ambiguities to arise. This work proposes a solution for such localization ambiguities. The proposed method assumes that the position of acoustically-tagged marine animals follows a hidden Markov model, such that localization ambiguities can probabilistically be resolved using a Forward-Backward algorithm. Our method is able to extrapolate the positions in a data series, as long as one sample in that series is picked up by three receivers, or if the identity of the receivers changes during tracking. Performance analysis shows that the localization accuracy of our method approaches the Cramér-Rao lower bound. Furthermore, to demonstrate the suitability of our method in a real sea environment, we have established a testbed that operated for three months, demonstrating localization of 20 acoustically-tagged sandbar sharks. Compared to the available solutions, roughly 20 times more location estimates were made; thereby, significantly increasing the impact of the test-site.

Internet-of-Underwater-Things (IoUT) systems comprising tens or hundreds of underwater acoustic communication nodes will become feasible in the near future. The development of scalable networking protocols is a key enabling technology for such IoUT systems, but this task is challenging due to the fundamental limitations of the underwater acoustic communication channel: extremely slow propagation and limited bandwidth. The aim of this article is to propose the JOIN protocol to enable the integration of new nodes into an existing IoUT network without the control overhead of typical state-of-the-art solutions. The proposed solution is based on the capability of a joining node to incorporate local topology and schedule information into a probabilistic model that allows it to choose when to join the network to minimize the expected number of collisions. The proposed approach is tested in numerical simulations and validated in two sea trials. The simulations show that the JOIN protocol achieves fast convergence to a collision-free solution, fast network adaptation to new nodes, and negligible network disruption due to collisions caused by a joining node. The sea trials demonstrate the practical feasibility of this protocol in real underwater acoustic network deployments and provide valuable insight for future work on the tradeoff between control overhead and reliability of the JOIN protocol in a harsh acoustic communication environment.

With the massive development of underwater small robotic vehicles and matching acoustic modems, applications for the Internet of Underwater Things (IoUT) are emerging. IoUT involves communication between the nonsynchronized network nodes organized in a mesh. A limiting factor of such communication is the so-called near-far effect, where transmissions from a node (near) close to a common receiver block the transmissions of a farther node (far). Due to the high-power attenuation in the underwater acoustic channel, near-far effect is common in underwater acoustic communication networks, and the phenomena occurs even for a distance ratio of 80% between the near and far nodes to the receiver, and the large number of nodes in the IoUT compounds the effect of this phenomena. While the current approaches only consider the jamming effect to the far signal, in this article, we consider canceling the interference from both sources by estimating and equalizing the channels on parallel, thereby significantly improving the decoding of both signals. As a result, IoUT performance improves. To limit mutual interference, we propose an automatic switching mechanism that controls the cancelation operation both in channel estimation and channel equalization. The simulation results show that our approach obtains significant improvement in communication from both near and far nodes. Results from a designated sea trial demonstrate that when both nodes are affected by their mutual transmissions, our proposed method improves the output signal-to-noise ratio (SNR) significantly.

The recent boost in undersea operations has led to the development of high-resolution sonar systems mounted on autonomous vehicles. These vehicles are used to scan the seafloor in search of different objects such as sunken ships, archaeological sites, and submerged mines. An important part of the detection operation is the segmentation of sonar images, where the object's highlight and shadow are distinguished from the seabed background. In this work, we focus on the automatic segmentation of sonar images. We present our enhanced fuzzybased with Kernel metric (EnFK) algorithm for the segmentation of sonar images which, in an attempt to improve segmentation accuracy, introduces two new fuzzy terms of local spatial and statistical information. Our algorithm includes a preliminary de-noising algorithm which, together with the original image, feeds into the segmentation procedure to avoid trapping to local minima and to improve convergence. The result is a segmentation procedure that specifically suits the intensity inhomogeneity and the complex seabed texture of sonar images. We tested our approach using simulated images, real sonar images, and sonar images that we created in two different sea experiments, using multibeam sonar and synthetic aperture sonar. The results show accurate segmentation performance that is far beyond the stateof-the-art results.

Long-range underwater acoustic communication (LR-UWAC) refers to the peer-to-peer transmission of messages for distances of tens to hundreds of km. It is a key enabling technique for applications such as control over unmanned underwater vehicles for long-term surveying. While underwater acoustic communication over shorter ranges is an established technique, this is not the case for LR-UWAC. This gap is mostly due to channel uncertainties: in the absence of feedback from the receiver and due to the long transmission range, channel state information (CSI) at the transmitter may not reflect the actual channel. In this paper, we propose an adaptive approach to pre-set the modulation scheme for LR-UWAC. This is a channel classification approach which, based on environmental information and on prior training on various channel types, predicts the best modulation scheme for the expected channel. Our classification procedure is trained to identify the channel's important features. Thus, compared to a direct decision approach, it becomes less sensitive to possible mismatches of environmental information. Our numerical simulation and sea experiment show that our approach successfully identifies the best modulation scheme based on the environmental information - even when the information is biased or only partially available.

We address the challenge of detecting an arbitrary-shaped underwater acoustic signal. Instead of setting a detection threshold, which due to noise transients may result in a high false alarm rate (FAR), our method classifies each measured sample as either “noise” or “signal.” Utilizing a priori knowledge of only the minimal duration of the signal, the decision is made using loopy belief propagation over a factor graph. Numerical simulations and sea experimental results show that our scheme achieves a favorable tradeoff between the Recall and FAR, and noise robustness, which far exceeds that of benchmark schemes.

We present a non-linear navigation solution, referred to as the Reverse Bearing Only Target Motion Analysis (Reverse BO-TMA). Reverse BO-TMA is a passive method for the self-localization of an Autonomous Underwater Vehicle (AUV). Our method relies solely on bearing measurements based on the radiated noise of a passing vessel sailing along a known route. Compared to traditional range-based underwater localization methods, Reverse BO-TMA allows the AUV to remain farther from the reference vessel, and does not require collaboration or message exchange. We formalize the Reverse BO-TMA as an optimization problem, and offer both a least squares solution and an unscented Kalman filter solution. Numerical results show that Reverse BO-TMA provides accurate performance, in terms of both positioning and speed, which are close to the posterior Cramer-Rao lower bound. To demonstrate the effectiveness of our approach, we have implemented a prototype for Reverse BO-TMA and successfully tested it in three sea experiments. We show that the Reverse BO-TMA is suitable for the long-term deployment of an AUV and in cases where energy is scarce and cooperating anchors are not available.

With the growing use of underwater acoustic communications (UWAC) for both industrial and military operations, there is a need to ensure communication security. A particular challenge is represented by underwater acoustic networks (UWANs), which are often left unattended over long periods of time. Currently, due to the physical and performance limitations, the UWAC packets rarely include encryption, leaving the UWAN exposed to external attacks faking legitimate messages. In this paper, we propose a new algorithm for message authentication in a UWAN setting. We begin by observing that, due to the strong spatial dependency of the underwater acoustic channel, an attacker can attempt to mimic the channel associated with the legitimate transmitter only for a small set of receivers, typically just for a single one. Taking this into account, our scheme relies on trusted nodes that independently help a sink node in the authentication process. For each incoming packet, the sink fuses beliefs evaluated by the trusted nodes to reach an authentication decision. These beliefs are based on the estimated statistics of the channel parameters, which are chosen to be the most sensitive to the transmitter-receiver displacement. Our simulation results show accurate identification of an attacker's packet. We also report results from a sea experiment demonstrating the effectiveness of our approach.

We focus on the detection of sporadic low-power acoustic/seismic signals of unknown structure and statistics, such as the detection of sound produced by marine mammals, low-power underground signals, or the discovery of events such as volcano eruptions. In these cases, since the ambient noise may be fast time varying and may include many noise transients, threshold-based detection may lead to a significant false alarm rate. Instead, we propose a detection scheme that avoids the use of a decision threshold. Our method is based on clustering the samples of the observed buffer according to a binary hidden Markov model to discriminate between “noise” and “signal” states. Our detector is a modification of the Baum-Welch algorithm that takes into account the expected continuity of the desired signal and obtains a robust detection using the complex but flexible general Gaussian mixture model. The result is a combination of a constrained expectation-maximization algorithm with the Viterbi algorithm. We evaluate the performance of our scheme in numerical simulations, in a seimic test, and in an ocean experiment. The results are close to the hybrid Cramér-Rao lower bound and show that, at the cost of some additional complexity, our proposed algorithm outperforms common benchmark methods in terms of detection and false alarm rates, and also achieves a better accuracy of the time of detection. To allow reproducibility of the results, we publish our code.

While acoustic communications are still considered the most prominent technology to communicate under water, other technologies are being developed based, e.g., on optical and radio-frequency electromagnetic waves. Each technology has its own advantages and drawbacks: for example, acoustic signals achieve long communication ranges at order-of-kbit/s rates, whereas optical signals offer order-of-Mbit/s transmission rates, but only over short ranges. Such diversity can be leveraged by multimodal systems, which integrate different technologies and provide the intelligence required to decide which one should be used at any given time. In this paper, we address a fundamental part of this intelligence by proposing optimal multimodal routing (OMR), a novel routing protocol for underwater networks of multimodal nodes. OMR makes distributed decisions about the flow in each link and over each technology, at any given time, in order to advance a packet toward its destination; in doing so, it prevents bottlenecks and allocates resources fairly to different nodes. We analyze the performance of OMR via simulations and in a field experiment. The results show that OMR successfully leverages all technologies to deliver data, even in the presence of imperfect topology information. To permit the reproduction of our results, we share our simulation code.

We present a spatial reuse resource allocation scheme for underwater acoustic networks that organizes communications so as to avoid destructive collisions. One prime source of collisions in underwater acoustic networks is the so called near-far effect, where a node located farther from the receiver is jammed by a closer node. While common practice considers such a situation a challenge, in this paper we consider it a resource, and use it to increase the network throughput of spatial-reuse time-division multiple access. Our algorithm serves two types of communications: (1) contention-free and (2) opportunistic. Our objective is to maximize the time slot allocation, while guaranteeing a minimum per-node packet transmission rate. The result is an increase in the number of contention-free packets received, and a decrease in the scheduling delay of opportunistic packets. Numerical results show that, at a slight cost in terms of fairness, our scheduling solutions achieve higher throughput and lower transmission delay than benchmark spatial-reuse scheduling protocols. These results are verified in a field experiment conducted in the Garda Lake, Italy, where we demonstrated our solution using off-the-shelf acoustic modems. To allow the reproducibility of our results, we publish the implementation of our proposed algorithm.

Wireless transmissions in water are mostly carried out via long-range (but low-rate) underwater acoustic communications, or short-range (but high-rate) underwater optical communications. In this paper, we are interested in finding out whether a statistical relationship exists between underwater acoustics and optics. Besides the theoretical interest of such a relationship, predicting the quality of the optical link through acoustics is also relevant in the context of a multimodal system with both acoustics and optics. Our study is based on a large data set acquired during the NATO ALOMEX 2015 expedition. During this experiment, we simultaneously measured several characteristics of the acoustic and optical links at multiple locations, reflecting a diversity of sea environments. Our results show a strong correlation between the properties of the acoustic link and the reliability of optical communications. This correlation makes it possible to predict the state of the underwater optical link at a certain depth and range. Due to the complexity of the acoustic and optical channels, we could not find the source of this correlation. This paper is, therefore, aimed to stimulate a theoretical study of the mutual properties of underwater acoustic and optical communication links. For reproducibility, we share the processed data from the experiment.

The low sound speed in water makes propagation delay (PD)-based range estimation attractive for underwater acoustic localization (UWAL). However, due to the long channel impulse response and the existence of reflectors, PD-based UWAL suffers from significant degradation when PD measurements of nonline-of-sight (NLOS) communication links are falsely identified as line-of-sight (LOS). In this paper, we utilize expected variation of PD measurements due to mobility of nodes and present an algorithm to classify the former into LOS and NLOS links. First, by comparing signal strength-based and PD-based range measurements, we identify object-related NLOS (ONLOS) links, where signals are reflected from objects with high reflection loss, for example, ships hull, docks, rocks and so on. In the second step, excluding PD measurements related to ONLOS links, we use a constrained expectation-maximization algorithm to classify PD measurements into two classes: LOS and sea-related NLOS (SNLOS), and to estimate the statistical parameters of each class. Since our classifier relies on models for the underwater acoustic channel, which are often simplified, alongside simulation results, we validate the performance of our classifier based on measurements from three sea trials. Both our simulation and sea trial results demonstrate a high detection rate of ONLOS links, and accurate classification of PD measurements into LOS and SNLOS.

Underwater acoustic localization (UWAL) is a key element in most underwater communication applications. The absence of GPS as well as the signal propagation environment makes UWAL similar to indoor localization. However, UWAL poses additional challenges. The propagation speed varies with depth, temperature, and salinity, anchor and unlocalized (UL) nodes cannot be assumed time-synchronized, and nodes are constantly moving due to ocean currents or self-motion. Taking these specific features of UWAL into account, in this paper, we describe a new sequential algorithm for joint time-synchronization and localization for underwater networks. The algorithm is based on packet exchanges between anchor and UL nodes, makes use of directional navigation systems employed in nodes to obtain accurate short-term motion estimates, and exploits the permanent motion of nodes. Our solution also allows self-evaluation of the localization accuracy. Using simulations, we compare our algorithm to two benchmark localization methods as well as to the Cramér-Rao bound (CBR). The results demonstrate that our algorithm achieves accurate localization using only two anchor nodes and outperforms the benchmark schemes when node synchronization and knowledge of propagation speed are not available. Moreover, we report results of a sea trial where we validated our algorithm in open sea.

In this paper we address the problem of spatialreuse scheduling for underwater acoustic communication (UWAC) networks that support high traffic broadcast communication and require robustness to inaccurate topology information. To this end, we derive a broadcast scheduling algorithm that combines topology-transparent and topology-dependent scheduling methodologies to achieve high-throughput in static and dynamic topology scenarios. While we focus on scheduling in UWAC networks, our approach can also be adopted for broadcast scheduling for radio ad-hoc network if robustness to topology uncertainties is desired. Simulation results for typical UWAC scenarios demonstrate that our protocol achieves high throughput and provides robustness to outdated topology information in dynamic topologies.