M. H. Supriya

Discrete Sine Transform based HMM underwater signal classifier

Underwater target recognition and classification has been a field of considerable importance due to its multidimensional applications. Much attention has been focused on this area and various underwater signal processing schemes have been devised over the time. Hidden Markov Models, because of their robustness, provide an effective architecture for the classification of underwater noise sources. A methodology is presented, in this paper, for the design and performance analysis of an HMM based underwater signal classification system, utilizing the Discrete Sine Transform based target specific features. Simulation results have been presented for typical underwater noise waveforms, such as Boat and Dolphin noises.

Localisation of underwater targets using sensor networks

Precise localisation and tracking of underwater targets in regimes like oceanography, fisheries and military applications is of prime importance owing to various phenomena that impede the accuracy of physical measurements. An approach for localisation of unknown underwater targets using a minimally configurable, three buoy sensor networks using passive listening concepts has been prototyped. The system comprises three sensor nodes, each consisting of mechanically steerable hydrophone array and support electronics. The sensor system picks up the noise emanations from the targets of interest and the hydrophone array of each node gets aligned to the Direction of maximum signal Arrival DOA. Using the DOAs measured, the distances of the target from the three nodes are computed. The prototype localiser model has been field tested and the validation tests yielded encouraging results within the limits of theoretical approximations and measurement errors, which necessitated the need for refining the estimates using Kalman filter.

Cost effective sensor buoy for ocean environmental monitoring

A low cost prototype sensor module comprising of a microcontroller and digital sensors for measuring the temperature, light as well as pressure, housed in a miniaturized buoy similar to a Sonobuoy has been developed for sampling the sea surface temperature, light intensity and sea level changes respectively. This standalone data logger module is interfaced with a ZigBee module, allowing addressable secure communication, for real time transfer of the parameters to a shore station, which will have a very good correlation with climate. The sensor module yielded satisfactory results during the field trials carried out in a hydroelectric reservoir with a small reduction in the measured values due to the encapsulation. An RF link was also established between the buoy and the shorestation and its performance was evaluated.

Deep learning architectures for underwater target recognition

Passive sonar target recognition is a challenging task due to the complex milieu of the ocean. Most of the state of the art target recognition systems depend on hand engineered feature extraction schemes in order to effectively represent the target signatures, based on expert knowledge. Due to the whimsical nature of the sources and medium, such feature engineering methods often fail to yield invariant features from the observations. In this paper, a deep unsupervised feature learning approach capable of capturing invariant features from the sensory signal stream through multi layered hierarchical abstraction has been adopted. These abstractions learned by the higher layers are mostly invariant and can be used as the discriminative features for the purpose of classification.

Tracking of a maneuvering underwater target

This paper elaborates a technique for improving the performance of a sensor network based system for tracking an abruptly maneuvering under water target. The results of tracking estimates of a maneuvering target may vary owing to various noises and interferences such as sensor errors and environmental noises. The conventional Kalman filter may induce unsatisfactory tracking errors when applied to the maneuvering target scenario, since the parameters of the filter cannot adapt itself to the highly maneuvering target. In this simulation study, a decision based maneuvering detection which depends on the chi-square significance test of the measurement residuals has been exercised. Upon detection of the maneuvering, the Kalman filter is reinitialized by resetting the parameters for improving the maneuvering target tracking estimates.

Improving the localization estimates using Kalman filters

Localization of underwater targets is an important requirement in surveillance operations. A method for improving the accuracy of the estimated location of an unknown target using Kalman filters is presented in this paper. The localisation is carried out using a sensor network comprising of three fixed surface buoy systems, each consisting of steerable hydrophone arrays and support electronics, positioned at the vertices of a triangle. The system picks up the noise emanations from the targets of interest and the hydrophone arrays of each node gets aligned to the direction of maximum signal arrival (DOA). Using the DOAs measured at each buoy system, the distances of the target from the three nodes are computed using the triangulation technique.

Prototype archival tags for studying the migratory routes of tuna

Archival electronic tags can be used for studying the migratory patterns of tuna, identifying their feeding and spawning grounds, etc. The size of the tag has to be miniaturized, to the extent possible so that by way of attaching such devices the swimming and natural behaviour of the tagged species remain unaffected. The design and development prototype archival tags for sampling the physical parameters of the ocean such as pressure, temperature and light intensity at preset intervals of time is presented in this paper. The tag has been encapsulated and its performance validated in the laboratory as well as in an open test facility in a reservoir.

Development of a hardware based underwater target identification system

An underwater target identification system comprising of a sensing element signal conditioner and a digital signal processor used for identifying targets of interest is presented in this paper. The sensing element picks up the noise waveforms emanating from the targets and the signal conditioner pre-processes and filters the captured signal into the desired spectral band for signal analysis. Signature features of the captured signal are extracted using the power spectral statistics implemented in a digital signal processing hardware. The features extracted are then compared with those available in the knowledge base, using pattern matching techniques, leading to the identification of underwater targets.

Underwater image enhancement using CLAHE in a reconfigurable platform

The main idea in adaptive histogram equalization is to find the mapping for each pixel based on it local (neighborhood) gray scale distribution. In this method the contrast enhancement mapping applied to a particular pixel is a function of the intensity values of pixels immediately surrounding the pixel. Hence the number of times that this calculation should be repeated is the same as the number of pixels in the image. This gives rise to an extensive computation requirement, which even with some modifications cannot be used for real time image enhancement. So here another form of adaptive histogram equalization is used which is a compromise between global histogram equalization and fully adaptive histogram equalization - regional histogram equalization. The noise is also suppressed by contrast limited enhancement. The design is implemented in Field Programmable Gate Array (FPGA) which is known to be a better choice for hardware implementation where parallel processing algorithms such as image processing is carried out. The algorithm is successfully implemented in Xilinx Spartan 3AN on Altium Nanoboard NB3000 board using Altium Designer.

Underwater target classifier using modified Kaiser-Bessel window

Underwater target classification has got numerous applications in ocean systems and technologies. The selection of suitable source specific features in a classifier system is one of the major factors determining the efficiency and efficacy of the classifier. The spectral features, when suitably modified, can provide certain essential clues suitable for the design of underwater signal classifiers. In this paper, a non-stochastic underwater target classifier, making use of an efficient feature set based on modified Kaiser-Bessel window, operating in the frequency domain is proposed. The proposed classifier is making use of a Matching Parameter which is a functional measure of the Mahalanobis and Euclidean distances and utilizes an algorithmic vector quantization approach as well for cluster formation. The system performance has been studied and fairly acceptable success rates have been obtained for the proposed underwater target classifier, making use of a modified Kaiser-Bessel window.