Present day sonars uses hundreds of sensors to make big arrays. These sensors are often digitized separately and the beamformer is realized in digital data. Current processor technologies support enough computations. The processors can be effectively utilized only if the memory and cache can be managed well. This paper analyses a time domain beamformer and the cache related issues are studied. We evaluate the use of techniques specific to processor architecture as well as some general signal processing technique to achieve high-performance digital signal processing on general-purpose processors. The results will be presented towards the end of the paper.

The matched filter output performance of wave forms affected by delay and Doppler are often analyzed by narrowband ambiguity functions in RADAR and radio communication literature. The same ideas are often adopted in the active SONAR and Underwater Wireless Acoustic (UWA) communication designs. UWA channels are wide band in nature and hence many of the simplifying assumptions commonly used for narrowband channels may not be applicable. Different types FM waveforms are analyzed using narrowband and wide band definition of ambiguity function to illustrate the effects of Doppler in narrowband and wide band channels on these wave forms.

Highly accurate estimation of Direction of Arrival (DOA) of underwater targets with less number of hydrophones has always remained a challenging problem among the acoustic signal processing research community. In addition to the recurring overheads in terms of money, the large footprints of the conventionally used arrays with hundreds of sensors make them infeasible to be used along with small platforms. Motivated by the hearing ability of animals with two ears, this paper proposes a two-sensor DOA estimation approach utilizing the binaural cues extracted from the target signals of interest. An unsupervised GMM clustering is performed on the extracted features in order to estimate the DOA of targets within an angular range of 0° to 180°. Evaluation of the proposed system using various performance metrics yields promising results on binaural data synthesized from underwater target records acquired during various expeditions conducted in the Indian Ocean.

Lead Zirconate Titanate (PbZrTiO3) (PZT) has been used for underwater acoustic sensing in different and acoustically band-limited. Varying sizes of them are needed for their operation in varying applications and frequency regimes. Therefore, their utilization is limited especially to the recent ocean-going smart, unmanned and programmed vehicles like autonomous underwater vehicles (AUV), remotelv operated vehicles (ROV), bio-mimicking underwater vehicles (BUV), etc. Thus, research on transforming the conventional bulk form of PZT to the thin-film form is on the rise. This paper attempts to report some of the advances on PZT thin films that are evolving towards the realization of underwater acoustic sensors.

Demand for miniature hydrophone sensors is ever increasing due to the advent of smart miniature autonomous underwater vehicles. Piezo micro-electro-mechanical system (MEMS) based devices are being reported for acoustic sensing applications. In this work, stainless steel (SS) and fibre reinforced plastics (FRP) materials are used as the packaging base for the piezo MEMS structure. Analysis of Transducers by Integrating Laplace equations (ATILA) and GiD are used to perform the analysis of the packaged device and receiving sensitivities for different dimensions of the piezo MEMS. The effect of materials and their thickness are simulated and reported.

We investigate underwater acoustic environment characteristics and shallow water channel interpretation through variation of the multipath scattering parameters and their impact on the Delay-Doppler spread function using geometric feature representations. We generate a diverse portfolio of simulated channels using well-known statistical models of shallow water acoustic multipath propagation, across variations of shallow underwater acoustic channel parameters. Specifically, we explore four different variables: water depth, channel range, spreading factor, and number of multipath arrivals, while following the variations shown on delay-doppler spread function. In each simulated channel, we provide the feature representation of the channel interpreted in terms of the separable channel multipath arrivals. This work can potentially be beneficial in understanding distinct underwater terrains in real time as well as predicting the rapid fluctuations in the shallow water channel, both in delay-time and delay-doppler domain representations.

Acoustic propagation in shallow water is greatly influenced by the properties of the sea bottom. Most of the continental shelf regions of the world have a surficial sediment layer consisting of granular materials such as sand and silt. This is especially true for the east coast of the United States. There has been significant amount of research conducted over the past decades understanding the acoustic properties (compressional and shear wave speeds and attenuations) of these sediments and its impact on acoustic propagation. Compared to granular (sandy) sediments, much less is known about the acoustics of fine-grained sediments such as clay and mud. There has been limited studies undertaken in the past to understand the impact of a low velocity surficial sediment layer on acoustic propagation. This study demonstrates the possibility of trapped modes in the surficial soft sediment layer based on model results. Effect of some of the environmental parameters such as the changes in the sound speed in water, thickness of mud layer, changes in the sound speed in the acoustic basement is investigated.

Target signal detection in the presence of tow-ship noise is an important requirement in towed array sonar applications. The self-interference is a severe problem especially in the ocean conditions where multipath propagation is supported. In this paper, we propose a novel deep learning-based algorithm using deep regression neural network (DRNN) for effective self- noise suppression with minimal power loss for target signal. In the training phase, the designed DRNN tries to minimize the root mean square error between the received signal and the desired output which is the target alone signal. In the testing phase, the DRNN is fed with the received signal and the neural network learns to suppress the self-noise caused by the tow ship. Due to the self-learning ability and high feature extraction capability, the network performs extremely well. To the best of our knowledge, deep learning-based algorithm is employed for the first time, for tow-ship noise suppression application, as per literature. The efficacy of the proposed method is evaluated over simulated data and the simulations were done as received by a towed linear array of acoustic sensors in the shallow ocean scenario. Experimental results show that the proposed method can improve the accuracy significantly.

In this work, we put forward the approach of fusion of two-color corrected images of the same single underwater image obtained by two different color correction methods to optimize the restoration of colors of the image. Also, selective color correction is followed instead of general color correction. By this approach, we can reduce the color casts, a common drawback of the existing enhancement methods. The quality of the result of our fusion method is compared with the existing two fusion-based enhancement methods of single underwater images. The experimental result shows that our method outperforms to compensate for the color distortions in underwater images.

Most of the applications with SONAR are relied on the Direction of Arrival (DOA) estimation. It is an inevitable part in the scenarios of target detection and localization where the location of target is retrieved from the received sensor data. The paper models novel deep learning architectures for DOA estimation and compares their performance with the existing DOA estimation algorithms like conventional beamforming, MVDR beamforming and MMSE estimation. The deep learning frameworks like Deep Neural Network (DNN) and Deep Convolutional Neural Network (DCNN) are deployed to recover the signal radiated from passive targets using multiple sensor arrays. The estimation is carried out for the SNR conditions ranging from -5dB to 20dB and the results are analyzed using MSE v/s SNR plot, waterfall diagrams and the time taken for execution.

Enhancing degraded underwater images is a challenging problem for researchers since such images have important applications in diverse research areas. Deep learning approaches contribute a significant improvement in image restoration. Automatic analysis of marine digital data has a higher research priority. This paper aims to introduce a fully connected convolutional neural network for dehazing underwater images. The deep encoder and decoder framework combine both low-level and high-level features which helps to recover the hazy image. Experiments evaluated on a benchmark dataset shows that the proposed method outperforms many state-of-the-art methods in terms of quantitative and visual quality. The network has the ability to dehaze underwater images while preserving fine details.

Deep convolutional neural networks-based architectures have led to rapid development in the field of segmentation. Image segmentation has been increasingly enforced in many areas, for example robotic vision, navigation, virtual reality and augmented reality. Research are also being made in underwater images to explore vast marine biological resources and gene banks. This paper tries to comparative performance analysis of image segmentation based on unet architecture for underwater object detection by semantic image segmentation. The proposed model is evaluated on five species of Fish4Knowledge image dataset. By choosing appropriate hyperparameters good segmentation results achieved for threshold values in between 0.5 and 0.6.

In this paper, the Doppler Velocity Log (DVL) is integrated with ArduSub Firmware. The ArduSub firmware is a fully-featured open-source solution for Remotely Operated underwater Vehicles (ROVs) and Autonomous Underwater Vehicles (AUVs). An attempt is also made to use SITL, which is a simulator that allows one to run ArduSub without hardware.

The heterodyne correlator detector is a computationally efficient implementation of the matched filter detector for narrow band active sonar waveforms. It is the common choice for signal detection systems when real time requirements are to be met with limited hardware resources. This paper presents an analysis of the performance bounds of the heterodyne correlator as compared to optimum match filtering. The detection threshold of a heterodyne correlator detector is arrived at from the analysis and validated by simulations. The losses due to the signal overlap mismatch which is inevitable in practical implementation of such systems are also presented.

The acoustic channel in a shallow ocean condition can be modeled as a doubly dispersive channel with significant Doppler and delay spread. Commonly used detector in active sonar system is known as replica correlator. The relative movements between own platform and interested target introduce Doppler shift in the received target echo. In such scenario, a bank of correlator with different Doppler shifted replicas is implemented in active sonar signal processor and the maximum value among different correlator output is chosen as the detector output. The plane wave assumption does not hold in shallow water condition with wide-band active transmitted signal. Moreover, the Doppler shift introduced in each ray path depends on the arrival and departure angles at the transmitter and receiver side. In this paper, we derive a generalized likelihood ratio test (GLRT) detector for the doubly dispersive underwater acoustic channel, which can utilize the multiple arrivals to enhance the detection performance. Detection threshold (DT) of a detector is an important parameter involved in the design of the sonar figure of merit (FOM). We derive the closed form expression of the detection threshold of introduced detector and compare it with other detectors. The derived expression shows that the performance of the detector for shallow water waveguide depends on the l2 norm of the channel vector.

Ocean surveillance using sonar uses a panoramic display for a 360◦ undersea investigation. Beamformers have been widely used for this purpose. They are designed to filter out noises and other incoherent signals from all directions while enhancing the signal-to-noise ratio from the desired direction essentially to detect targets and estimate their location. Sonar arrays formed by several sensing elements arranged in different shapes are used for forming pre- formed or steered beams in all directions. Such pre-formed beams are expected to be identical and unperturbed. Non-identical beams result in errors in bearing, resolution and discrimination, etc. Beamwidth, sidelobe levels (SLL), beam sensitivity and tolerance to the array integrity are the deciding factors of beam identicality. They can be controlled by providing optimized individual weighting called windowing. These optimizations are often complex and sensitive to mismatches in individual sensors, namely variations in receiving sensitivity, error in sensor location, failure of sensors, and acoustic centre, etc. Reliability analysis of a novel product beamformer is made in this paper against the seven most commonly used standard beamformers to choose a reliable beamformer. Failure in the array element is the criteria set for the analysis. The presented results suggest that the product beamformer is more stable than traditional beamformers.

In a sonar system, the audio time series is an important tool for the sonar operator to validate and classify the target identified from the sonar display. The audio time series is also useful to confirm the target in the presence of reverberation. The beam time series is given as the input to the audio streamer. Comparing with the sensor level signal, the audio time series has an SNR (signal to noise ratio) improvement due to the directivity index (DI) of the beamforming. The beam time series is further processed using the matched filter and the normalized detector output is shown in the sonar display. Hence, the detector output in the sonar display has the gain of beamformer as well as matched filtering. However, the SNR of the audio time series is not enhanced by the detector gain. Hence in a scenario with weak SNR target, the target can be identified in the sonar display and the operator fails to confirm the target in the sonar audio. We introduce a subspace based algorithm to enhance the SNR of the sonar audio signal. The simulation as well as experimental results establish the superior performance of the proposed algorithm to improve the SNR of the audio time series significantly.

System Integration of a Sonar suite consisting of multiple Sonar systems results in unforeseen errors which if not localized at the source lead to increase in system development cost and time. Test and evaluation equipment are required for fault localization during each phase of system integration tests and for overall system performance and evaluation. This paper describes the design of a virtual instrument with multifunctional test capabilities implemented in a desktop PC and its usage to carry out system integration tests. The instrument functionalities are realized as software solution using built in multithreading, multi-tasking features in graphical programming environment of LabVIEW.

Detection, localization, tracking and classification of underwater targets in the ocean assume great significance and attract immense attention due to its growing importance in various spheres, especially for military applications. Several sonar systems have been established for this purpose over the years and practiced by researchers across the globe. Most of such systems use the methods of underwater acoustics as the customary standard. However, the advances in stealth technology and the complexity involved in ocean acoustics prevent the evolution of acoustic technologies and thus motivates the researchers to examine alternate techniques for the detection of underwater targets in the ocean. Different non-acoustic detection methods are suggested in the recent past to overcome the stealth and acoustic complexity related obstacles. However, the research in this area is scarce and many advances are expected to be originated. This paper investigates to use of some of the non-acoustic methods to detect the kind of non-acoustic signals namely hydrodynamic signatures such as wakes and internal waves by satellite image processing and advanced optical techniques.

Underwater noise statistics is an important input for the design of SONAR. A directional noise measurement and recording unit has been designed and developed at NPOL for deployment in deep waters up to 400 m. In this paper, the noise statistics of the measured directional and omnidirectional noise at three stations in deep water off Kochi is presented. The variation in noise level spectrum at different locations is qualitatively studied as a function of the measured wind speed and observed marine traffic. Also, omnidirectional noise level spectrum at different depths has been compared with the standard Wenz curves.

A detailed survey of research work in the area of fiber sensors has been carried out. The paper aims at the optimized design and development of air-backed concentric composite mandrel hydrophone to be used in under water sonar applications. Mandrel hydrophones with foaming layer over the solid base layer showed very good improvement in the sensitivity. For the use of underwater applications structure stability is required and hence a composite concentric air backed mandrel with a central Aluminium thin walled tube as base layer and the foaming layer of plastic syntactic foam above the base layer of 5mm thickness has been designed and fabricated. Three types of sensors with various outer diameter keeping the foaming layer thickness same, has been fabricated, after undergoing both tensile and compressive load tests of the material. These three sensors were put in an acoustic water tank and carried out the sensitivity measurements using a Michelson interferometric test set up. Encouraging results were obtained and shown.

A new fiber optic (FO) hydrophone, and an array configuration based on this new FO hydrophone suitable for submarine flank arrays is discussed. This FO hydrophone dimension is sufficiently large in size having the following significant features: 1) a planar hollow shape with curved side ends supporting smooth winding, and having a thin central metallic structure in the hollow shell, providing structural stability, leaving large cavity space on either side of the central structure, 2) easy component housing (delay coils, couplers, reflectors etc.) in the vacant cavity space, 3) single layer very long fiber winding, improving the acoustic sensitivity, 4) stiffening effect reduction of fiber on the mandrel, catering hydrophone with very long life, 5) a novel compact flank array realization only with hydrophones and fiber cable (all other components are accommodated inside the large hydrophone). The salient feature of this flank array is the accommodation of 64 sensors, on one wave length one fiber cable, by using a suitable multiplexing scheme. An optimized configuration has arrived with 512 sensors connected to 8 wavelengths, 8 fibers and meets the submarine flank array requirements. The 512 sensor system is repeated on the other side of the submarine thus meeting a very large flank array compatible with the larger submarines.

For active sonar operations, particularly airborne sonar, it is essential to transmit high power signals through a single-core coaxial cable. A projector is connected at the remote side for transmission of acoustic signal. Power loss in a coaxial cable varies logarithmically with the length. Single core coaxial cable introduces power losses due to its resistance, capacitance and inductance. If the cable is in wound condition, the inductance increases and majority of power lose is due to inductance. In order to deliver maximum power through the projector, proper tuning is required. In this paper, equivalent circuit model is implemented to analyse the effect of long cable on the performance of the acoustic transmitting system. Impedance, voltage transfer function and power across the effective resistive load of the transducer have been calculated and compared with simulation results.

This paper aims at providing a complete power amplifier solution, required for high frequency large transmitter arrays to meet collision avoidance sonar applications. It focuses on the design and development of : a) first of its kind approach for a fully digital power amplifier PCB of double euro standard, using eight APEX SA160A MIL-COTS switching power amplifier (PA) ICs, controlled by the MIL-COTS NUCLEO STM ARM M7 microcontroller board, (b) single standard 19 inch four level cabinet solution using 256 power amplifiers, by stacking eight PCBs (8 x 8 , 64 power amplifiers) in each level. c) master / slave network configuration (having single master and 32 slaves) through control area network (CAN) bus, for exchange of transmission parameters/health data between master slaves, and the Ethernet bus for communication between master with the host system d) novel hardware interrupt scheme controlled by the master, achieving highly precise synchronized power transmission, by simultaneously enabling all power amplifiers. The prototype testing and phase measurements with multiple PCBS were carried out using dummy loads and obtained highly accurate results. The phase accuracy of all channels are essential for accurate transmit beamforming and beam steering applications. This power amplifier array system can be used for any high frequency sonar transmitter array beam sweep operations.

Insulation resistance and capacitance are used as indicators of the health of hydrophones used for sonar applications on naval platforms. It is found that an increase in capacitance of the hydrophone indicates a very low value of insulation resistance. The reduction in insulation resistance and increase in capacitance leads to adverse effects in sonar systems, such as increased self-noise of the system and limited low- frequency operation. A very low value of insulation resistance indicates the end-of-life of the hydrophone. Sonar systems usually employ a charge amplifier for amplifying the hydrophone output. As the gain of the charge amplifier depends on the input capacitance, an increase in capacitance leads to an increase in preamplifier gain beyond the expected level, which may further lead to clipping of output signal. However, this opens a new way to test the health of the hydrophone without relying on external instruments. Determining the gain of the amplifier regularly using the reference transducer available in the platform helps determine the end-of-life on the hydrophone.

In the past few decades, need and thus interest in underwater voice communication research is growing due to its application in submarines, deep water manned submersibles, and diver’s communication. Compared to the initial Acoustic telephones, performance of the present systems is improved due to the continuous research interest. Very few review papers are available to provide the history of the development in this field. In this paper, it is attempted to provide an overview of the key developments over the past few decades with focus on voice communication in deep water manned submersible. Some of the open problems and challenges faced by the researchers are also discussed in this paper.