M. Girish Chandra

Complex Event Processing for object tracking and intrusion detection in Wireless Sensor Networks

Complex Event Processing (CEP) has received wider acceptability due to its systematic and multilevel architecture driven concept approach. CEP is an emerging technology in the field of data processing and identifying patterns of interest from multiple streams of events. High levels of integrated self learning applications can be developed. CEP is used in development of applications which have to deal with voluminous streams of incoming data with the task of finding meaningful events or patterns of events, and respond to the events of interest in real time. In this paper a CEP based application for object detection tracking in a Wireless Sensor Network (WSN) environment is proposed. Also the detection of an intruder using semantic query processing is proposed. ESPER, an open source Complex Event Processing engine is used to develop the application.

MobiCoStream: Real-time collaborative video upstream for Mobile Augmented Reality applications

MobiCoStream, an acronym for Mobile Collaborative upStream, refers to Real-time video upstream of captured video in Mobile Augmented Reality (MAR) applications. MAR for tele-assistance deals with providing real-time assistance to the user through the power of Augmented Reality. Real-time applications have strict delay requirements which need to be met in order to satisfy the user Quality of Service (QoS). Mobile communication still being in its third generation, limits the availability of high up-link bandwidth for video transmission. The problem is tackled by means of Bandwidth Aggregation through which the hand-held devices in the vicinity, collaborate with the user to achieve real-time video up-link and maintain the user QoS. In this paper, we present the actual results pertaining to the prototype implementation of the proposed system. Also presented are some simulation results, by viewing the multiple hand-held devices as a virtual Multiple Input Multiple Output (MIMO) system and in turn considering the popular Alamouti scheme in this context.

Security and Privacy for Real Time Video Streaming Using Hierarchical Inner Product Encryption Based Publish-Subscribe Architecture

In this era of digitization, there are many application swhich can be used for sharing videos to a group of people around the world in real time. During the acquisition of real time data, many sensitive information also gets captured along. This poses a threat to privacy of data. In this paper, we consider the scenario of real time sharing of video from multiple sources to multiple end users through a secure publish subscribe architecture. We propose a system that provides a secure user profile based data privacy using Hierarchical Inner Product Encryption (HIPE) and a broker with anonymous pub sub architecture. System provides multiple levels of data access control to end users based on their roles so that they will be able to see only those videos for which they have access. In order to ensure secure communication, the broker publishes data on encrypted topic and uses HIPE for data subscription. In this paper, we demonstrate the end-to-end working of the system through prototype implementation. Further, we also present the system performance and evaluate the system security both with and without HIPE.

Joint frequency and 2-D DOA recovery with sub-Nyquist difference space-time array

In this paper, joint frequency and 2-D direction of arrival (DOA) estimation at sub-Nyquist sampling rates of a multi-band signal (MBS) comprising of P disjoint narrowband signals is considered. Beginning with a standard uniform rectangular array (URA) consisting of M = Mx × My sensors, this paper proposes a simpler modification by adding a N — 1 delay channel network to only one of the sensor. A larger array is then formed by combining the sub-Nyquist sampled outputs of URA and the delay channel network, referred to as the difference space-time (DST) array. Towards estimating the joint frequency and 2-D DOA on this DST array, a new method utilizing the 3-D spatial smoothing for rank enhancement and a subspace algorithm based on ESPRIT is presented. Furthermore, it is shown that an ADC sampling frequency of fs ≥ B suffices, where B is the bandwidth of the narrow-band signal. With the proposed approach, it is shown that O(MN/4) frequencies and their 2-D DOAs can be estimated even when all frequencies alias to the same frequency due to sub-Nyquist sampling. Appropriate simulation results are also presented to corroborate these findings.

Data-driven electrical load disaggregation using graph signal processing

Graph Signal Processing is a very active area of research and is being increasingly used for different applications in different domains like vision, sensor networks etc. It deals with processing signals defined on the vertices of a graph which captures the relational dependencies of the signal. Recently, graph signal processing has been applied to source separation problem, in particular, electrical load disaggregation where, given smart meter measurements it is required to ascertain the contribution of multiple loads which could have resulted in those measurements. Inspired by these works, an attempt is made to adapt the results for real-life situations. Apart from incorporating pre-processing and post-processing stages, the disaggregation algorithm is appropriately modified to arrive at useful results for both load identification and consumption estimation. Both are important as disaggregation results, since the consumption information facilitates load prognostics/diagnostics and helps in giving advisory feedback to the consumers. Apart from outlining the algorithms, which work for different sampling rates like 10 seconds and 1 minute, some of the promising results obtained during the extensive simulation studies are presented in the paper.

Event and feature based electrical load disaggregation using graph signal processing

Electrical load disaggregation continues to attract new explorations due to its challenging nature as well as utility. When the loads to be separated are characterized by suitable features, there is a possibility to solve the problem by utilizing the techniques from the emerging area of Graph Signal Processing (GSP). In this paper, we propose a three-staged approach comprising of (i) Event Detection and Clustering (ii) Event Pairing and Feature Extraction and (iii) Load Classification, each of them being pivoted on GSP. For load classification in particular, a robust spectral clustering strategy is appropriately adopted using joint spectrum computed from different features. The efficacy of this novel combination is demonstrated through the results obtained on both public data sets and the simulated active power signals.

Joint frequency and direction of arrival estimation with space-time array

Joint frequency and DOA estimation of more sources than the number of sensors is considered in this paper. We assume a simple uniform linear array (ULA) and propose to employ a multiple delay channel network at every sensor, which can easily be realized by sampling at a slightly higher rate. By combining the outputs of the ULA and the delay network, we show that the manifold matrix is analogous to that of a uniform rectangular array, and hence we appropriately refer to this array as the space-time array. Further, estimation of parameters via the rotational invariance technique (ESPRIT) based algorithm referred to as space-time (ST)-Euler-ESPRIT is proposed. ST-Euler-ESPRIT, similar to well known unitary-ESPRIT provides automatically paired frequencies and their DOAs. We further show that with the proposed approach for a M element ULA and with N-1 delay channel, O(MN) frequencies and their DOAs can be estimated. The performance of ST-Euler-ESPRIT is verified by simulations, where it shows consistently better performance than the unitary-ESPRIT algorithm under noisy conditions.

Power system load data models and disaggregation based on sparse approximations

The deployment of smart meters by utilities holds the promise of improvements in operational efficiency, reliability and cost savings. With power measurements from smart meters, utilities can deploy innovative programs that allow end users to better control their energy usage while simultaneously reducing peak demand across the grid. In this paper, to develop data analysis tools for applications enabling monitoring and control of energy, a systems approach is taken, comprising of modeling, measurement, calibration and inference on the energy data collected from end users. A combination of analysis and synthesis for deriving data and measurement models calibrated to the aggregate power under measurement allows detection and estimation of features of individual appliances. Test results on disaggregation of power waveforms using the publicly available REDD data sets show promising results. The generic modeling and optimization framework can be used in the design and deployment of cyber physical energy systems for monitoring and control of energy resources.

Missing Data Interpolation Using Compressive Sensing: An Application for Sales Data Gathering

Tasks like survey analysis involve collection of large amounts of data from different sources. However, there are several situations where exhaustive data collection could be quite cumbersome or infeasible. In this paper, we propose a novel Compressive Sensing (CS)-based framework to recover the original data from less number of collected data points in the case of market or survey research. We utilize the historical data to establish sparsity of data, and further introduce the concept of logical proximity for better recovery results. Additionally, we also present a conceptual idea toward adaptive sampling using data stream sketching, which suggests whether the collected data measurements are sufficient or not. The proposed CS-based methodology is tested with toy-sized examples and the results are presented to demonstrate its utility.

A Real Time Surveillance System Using Wired and Wireless Sensor Networks by Multi-algorithmic Approach

In this paper, real time surveillance system is presented for an efficient identification of a person as an intruder or not in the inhibited location by multi-algorithmic approach. This system is aided with both wired and wireless sensor network. The Wireless Sensor Network detects the presence of person using PIR sensor and identifies the person with RFID and the verification is assured by using the image sensor associated with wired network. The data are gathered at server side and synchronised using time and location information. The image obtained is scanned for human faces and from the detected faces the features are extracted with the help of multiple algorithms (BICA, DCT&FLD, Kalman face) and the results are combined with weighted average based fusion strategy. The extracted features are then compared with the available set of features in database for authentication. The image comparison is done semantically by intra and inter class search in the database using the RFID value. The proposed system identifies a person as an intruder, if there is any mismatch in the sensed information. This system also provides solution to tail-gating problem. The system is implemented in JAVA with the OpenCV functions using JNA.