National Institute of Technology, Meghalaya

About: National Institute of Technology, Meghalaya is a education organization based out in Shillong, India. It is known for research contribution in the topics: Control theory & Electric power system. The organization has 503 authors who have published 1062 publications receiving 6818 citations. The organization is also known as: NIT Meghalaya & NITM.

Hyperspectral Remote Sensing Images and Supervised Feature Extraction

Abstract: In the last three decade, one of the significant breakthrough in remote sensing is to introduce of hyperspectral sensors, which acquire a set of images from hundreds of narrow and contiguous wavelengths of the electromagnetic spectrum from visible to infrared regions. Images, which are captured by these sensors, have detailed information in the spectral domain to identify and distinguish spectrally unique materials. To recognize the objects present in hyperspectral images, classification/clustering task need to be performed. But due to the presence of huge number of attributes, classification technique becomes more complex. So, before performing the classification task, reduce the number of attributes (denoted by dimensionality of the data) is an important step where the aim is to discard the redundant attributes and make it less time consuming for classification. In this chapter, few supervised feature extraction techniques for hyperspectral images i.e., prototype space feature extraction (PSFE), modified Fisher’s linear discriminant analysis (MFLDA), maximum margin criteria (MMC) based and partitioned MMC based methods are explained. Experiments are conducted over different hyperspectral data set with different quantitative measures to analyze the performance of these feature extraction methods.

Evolving Social Networks via Friend Recommendations

Abstract: A social network grows over a period of time withthe formation of new connections and relations. In recent yearswe have witnessed a massive growth of online social networkslike Facebook, Twitter etc. So it has become a problem ofextreme importance to know the destiny of these networks. Thuspredicting the evolution of a social network is a question ofextreme importance. A good model for evolution of a socialnetwork can help in understanding the properties responsiblefor the changes occurring in a network structure. In this paperwe propose such a model for evolution of social networks. Wemodel the social network as an undirected graph where nodesrepresent people and edges represent the friendship between them. We define the evolution process as a set of rules which resemblesvery closely to how a social network grows in real life. We simulatethe evolution process and show, how starting from an initialnetwork, a network evolves using this model. We also discuss howour model can be used to model various complex social networksother than online social networks like political networks, variousorganizations etc.

A Novel Edge-Supported Cost-Efficient Resource Management Approach for Smart Grid System

Abstract: The smart grids, a new-generation power supply system, have the capacity to lowering the cost, can increase service provision tremendously, and make surroundings greener as compared to conventional power supply systems. To interact with the physical world and widen its capabilities, integrated smart grid cyber-physical system (SG-CPS) can be used for computation, communication, and control. To support smart grid (SG), cloud components are employed for storing and processing users’ power demand and control flow information generated at different control components like smart meter (SM), home energy management (HEM), phasor measurement units (PMUs), and soon. But storing smart grid data to cloud and processing incurs unacceptable delays. This paper addresses quality-of-service (QoS) requirements of SGs by integrating fog computing along with cloud computing infrastructure for realizing an Edge Computing integrated Smart Grid (EC-iSG). To that end, this paper presents novel heuristics for resource management of such integrated infrastructure that accounts for parameters such as uplink and downlink communication costs, cost for VM deployment, and cost for communicating among base stations. The results presented demonstrate the efficacy of the proposed methodology.

Improvement in Performance of Induction Heating System Using Direct AC-AC Boost Converter

Abstract: Due to the unique advantages of induction heating (IH), it becomes the current technology of choice in the modern domestic application. So for designing of this system, efficiency is the key parameters which improve the performance of whole IH system. In order to achieve this goal, a direct AC-high frequency AC Boost converter is presented which converts utility AC to HFAC in a single stage conversion. Also, it reduces the component counts and stress level of voltage/current across the switches. This proposed topology is designed for 2.5 kW and verified through MATLAB/SIMULINK environment with extensive studies of simulation results.

Real-time analysis and simulation of multi-string grid connected photovoltaic inverter using FPGA

Abstract: The state-of-art in research practices across the globe emphasizes the real-time implementation of the control techniques. This paper investigates the performance of an Field Programmable Gate Array (FPGA)-based real-time implementation of a three-phase three-level Voltage Source Inverter (VSI) for grid integration of multi-string PV array. FPGA-based prototyping of control circuits enables the study of sensitivity on parameters variations, increases the safety and reduces the time and costs of implementation. The proposed system consists of a centralized multilevel inverter that aids in the conversion of DC power obtained from a multi-string PV array to AC power, which is fed to the utility grid as well as the three phase local loads connected at the Point of Common Coupling (PCC). The multilevel structure finds use in handling large amount of power, reducing voltage stress across the semi-conductor devices and reducing the harmonic distortion. The system is developed in Matlab/Simulink environment and the control algorithm of the VSI is implemented in Virtex-6 ML605 Evaluation-Kit with Xilinx System Generator providing the effective grid-interfacing environment.