Amir Ali Khan

A Source Separation Technique for Processing of Thermometric Data From Fiber-Optic DTS Measurements for Water Leakage Identification in Dikes

Distributed temperature sensors (DTSs) show real advantages over conventional temperature sensing technology such as low cost for long-range measurement, durability, stability, insensitivity to external perturbations, etc. They are particularly interesting for long-term health assessment of civil engineering structures such as dikes. In this paper, we address the problem of identification of leakage in dikes based on real thermometric data recorded by DTS. Formulating this task as a source separation problem, we propose a methodology based on Principal Component Analysis (PCA) and Independent Component Analysis (ICA). As the first PCA estimated source extracts an energetic subspace, other PCA sources allow to access the leakages. The energy of a leakage being very low compared to the entire data, a temporal windowing approach guarantees the presence of the leakages on these other PCA sources. However, on these sources, the leakages are not well separated from other factors like drains. An ICA processing, providing independent sources, is thus proposed to achieve better identification of the leakages. The study of different preprocessing steps such as normalization, spatial gradient, and transposition allows to propose a final scheme that represents a first step towards the automation of the leakage identification problem.

Automatic Monitoring System for Singularity Detection in Dikes By DTS Data Measurement

The development of automated monitoring systems for the detection of singularities, such as leakages in dikes, is indispensable to avoid mass disaster. An efficient solution for dike survey is the use of distributed temperature sensors (DTSs) based on optical fiber, offering a multitude of advantages such as low cost, extreme robustness, long-range measurement, etc. However, the temperature data acquired with DTSs, being not directly interpretable, require intervention of signal processing techniques. This paper addresses this signal processing aspect, exploiting the key idea that the temperature variations over the course of a day for singular zones are quite different from those for nonsingular zones. A daily reference temperature variation, which is representative of the nonsingular zones, is estimated using singular value decomposition (SVD). The residue subspace of SVD contains information linked to the deviations from this reference, thus allowing the degree of singularity to be quantified by a dissimilarity measure such as the L2-norm. To detect only the singularities in dikes, such as leakages or drains, a constant false alarm rate (CFAR) detector is proposed by modeling each daily dissimilarity measure with a mixture of Gamma and uniform distributions. The proposed automatic singularity detection system was validated under different scenarios on real data over periods from 2005 to 2007. The first scenario depicted the detection of percolation-type artificial leakages with their detection strength depending on their flow rates. Another scenario allowed detecting the presence of a real water leakage at the site, which was previously unobserved during manual inspections. The repeatability of the system was also verified by periodic analysis.

Inter comparison of classification techniques for vowel speech imagery using EEG sensors

The use of Electroencephalography (EEG) in the domain of Brain Computer Interface is a now common place. EEG for imagined speech reproduction and observation of brain response to audio stimuli are active areas of research. In this paper, we consider the case of imagined and mouthed non-audible speech recorded with EEG electrodes. We analyze different feature extraction techniques such as Mel Frequency Cepstral Coefficients (MFCCs), log variance Auto Regressive (AR) coefficients. Based on these extracted features, we perform a pairwise classification of vowels using three different classification models based on Support Vector Machine (SVM), Hidden Markov Models (HMM) and k-nn classifier. The proposed methodology is applied on four different data sets with some preprocessing techniques such as Common Spatial Pattern (CSP) filtering. The data sets principally comprised of either mouthing or solely imagining 5 vowel sounds without speaking or making any muscle movement. The goal of this study is to perform an inter comparison of different classification models and associated features for pairwise vowel imagery. The proposed approach is validated on different data sets and offer reasonable accuracies for pairwise classification.

Monitoring and early detection of internal erosion: Distributed sensing and processing

Early detection of leakages in hydraulic infrastructures is important to ensure their safety and security. Significant flow of water through the dike can be an indicator of internal erosion and results in a thermal anomaly. Temperature measurements are therefore capable of revealing information linked to leakage. Optical fiber–based distributed temperature sensors present an economically viable and reliable solution for recording spatio-temporal temperature data over long distances, with spatial and temperature resolutions of 1 m and 0.05°C, respectively. The acquired data are influenced by several factors, among them water leakages, heat transfer through the above soil depth, seasonal thermal variations, and the geomechanical environment. Soil properties such as permeability alter the acquired signal locally. This article presents leakage detection methods based on signal processing of the raw temperature data from optical fiber sensors. The first approach based on source separation identifies leakages by separating them from the non-relevant information. The second approach presents a potential alarm system based on the analysis of daily temperature variations. Successful detection results for simulated as well as real experimental setups of Electricité de France are presented.

SVD based automated dike monitoring system using DTS data

The detection of water leakages in dikes using distributed temperature sensors is an interesting prospect due to the commercial viability of these optical fiber based sensors. The acquired temperature data, being not directly interpretable, requires intervention of advanced signal processing techniques. In this work, we propose a system for the identification of singularities such as existing dike structures and water leakages. The distances where singularities exist show temperature variations over the course of a day which are different from the nonsingular zones. The different nonsingular zones though show a similar temperature variation trend. The proposed system estimates this reference trend as the most coherent component of the Singular Value Decomposition applied on daily data. The corresponding SVD residue subspace thus represents the deviation from the reference subspace and thus contains information on singularities. The L2 norm of this residue is a good discrimination measure for identification of these singularities.

Blind source separation techniques for percolation type leakage detection in dikes using fiber optic DTS signals

Distributed temperature sensors (DTS) based on fiber optics present an efficient means for temperature data acquisition. The use of DTS data to detect leakages in dikes necessitates some processing of this data. Formulating leakage detection as a source separation problem, the goal of this paper is to compare various blind source separation techniques for percolation type leakage detection using a real data set. Singular value decomposition can be used as the first step to separate out the ground response where acquisitions are made. Two independent component analysis algorithms, JADE and FastICA, assuming independence of sources, are tested to find the best solution for leakage detection.

Comparison of Classifier Architectures for Online Neural Spike Sorting

High-density, intracranial recordings from micro-electrode arrays need to undergo Spike Sorting in order to associate the recorded neuronal spikes to particular neurons. This involves spike detection, feature extraction, and classification. To reduce the data transmission and power requirements, on-chip real-time processing is becoming very popular. However, high computational resources are required for classifiers in on-chip spike-sorters, making scalability a great challenge. In this review paper, we analyze several popular classifiers to propose five new hardware architectures using the off-chip training with on-chip classification approach. These include support vector classification, fuzzy C-means classification, self-organizing maps classification, moving-centroid K-means classification, and Cosine distance classification. The performance of these architectures is analyzed in terms of accuracy and resource requirement. We establish that the neural networks based Self-Organizing Maps classifier offers the most viable solution. A spike sorter based on the Self-Organizing Maps classifier, requires only 7.83% of computational resources of the best-reported spike sorter, hierarchical adaptive means, while offering a 3% better accuracy at 7 dB SNR.

Mallat’s Scattering Transform Based Anomaly Sensing for Detection of Seizures in Scalp EEG

Epilepsy is one of the most common neurological disorders, which manifests as unprovoked seizures. The prevalence of epilepsy is higher in developing countries, where medical facilities are ill-equipped and under-staffed. Mobile EEG devices promise a new dawn for long-term ambulatory EEG monitoring, which has a potential to revolutionize health care for neurological disorders especially epilepsy. Increasing the outreach to underserved communities and continuous monitoring of patients will yield vast amount of data. This requires the development of a method that can mark regions of interest, to aid in the evaluation of the EEG trial by the experts. Such an experimental setting calls for an unsupervised method, which can detect seizure regions with high accuracy. This paper focuses on the development of a seizure detection method with the above-stated characteristics. Group invariant scattering, a novel data representation technique, has been used for feature extraction. Tested on CHB-MIT data set, the proposed methodology outperforms the current state-of-the-art approaches under similar testing conditions, by successfully detecting 180 out of 197 seizures.

Stereo-vision based autonomous underwater navigation — The platform SARSTION

Underwater robotics has been the center of attention for a long time due to diverse multi-disciplinary applications that it addresses. Conventionally, the detection and navigation has been governed by the SONAR based systems. While, SONAR offers unmatched capability for long range operation, its performance is compromised for detection of short range objects. In this regard, an alternate solution based on image processing appears promising in short range applications. Stereo-vision based systems are now regularly deployed for surface applications with huge success. However, the underwater deployment of these stereo-vision based systems is still an open problem. In the current work, we present an underwater autonomous platform, navigated by stereo-vision based algorithm. Initially, we demonstrate the distinct edge that the stereo-vision enjoys over SONAR in terms of its low cost, small size and low energy requirements. We present a working prototype of a Semi-Autonomous Sub-Aquatic Robot for Surveillance using Stereo-vision (SARSTION). SARSTION provides 6 degrees of freedom, using propellers, 4 for horizontal motion and 2 for vertical motion and is administered by an elaborate control system based on various positional feedback sensors. The main contribution of the current work is in the design of a novel simplistic navigational algorithm. The two acquired images from a front looking stereo-camera are used to compute a depth map whose intensity represents the relative distance of an object from the camera. The intensities are then equated to compute the relative physical distances of different objects from the camera. This information coupled with the approximations of relative object sizes and the data from the positional feedback sensors are then used to carefully alter the yaw of the mechanical assembly to maneuver the robot. The camera can also be used to provide live feed, depending upon its usage. We present some promising results illustrating the successful operation of the robot in a controlled reservoir.

A computationally efficient heart rate measurement system using video cameras

Video cameras are increasingly being used to measure human heart rates non-invasively, without contact. Such systems find applications in tele-medicine, remote monitoring of quarantined patients and premature neonates and are also useful for health conscious consumers. Several algorithms have been reported in the literature for measuring the heart rate from videos of human subjects. These algorithms use offline, computationally involved techniques such as Independent Component Analysis (ICA) or Principal Component Analysis (PCA) which render the algorithms unfeasible for implementation on real-time embedded systems. We conducted experiments to find an optimal colorspace for measuring the heart rate. We subsequently used a novel means of optical filtering of this colorspace to develop an accurate real-time algorithm, without the need for ICA or PCA, using ordinary standard definition (SD) web cameras. In this paper we present our algorithm and also compare it with existing state of the art algorithms.