Electronics

Vellore Institute of Technology (VIT) is a private university located in&nbsp;Tamil Nadu, India. Founded in 1984, as Vellore Engineering College, the institution offers 20 undergraduate, 34 postgraduate, four integrated and four research programs. It has campuses in Vellore, Amravati, Bhopal and Chennai.

VIT is one of the top ranked private universities in India according to NIRF, THE and QS Rankings.&nbsp;Govt. of India has recognized&nbsp;VIT, Vellore as an&nbsp;Institution of Eminence. This has allowed VIT to take independent quality initiatives and move up in world ranking.

&nbsp;

&nbsp;

VIT University

School of Electronics Engineering

The research work and applications of advanced materials have been increased in industries. At the same time, researchers are tried to convert the machining system into automatic process. In this paper is discussed about the internet of Things (IoT) based monitoring system of unconventional machining processes of Haynes 230 alloy. The water level, water pressure, abrasive quantity and its flow rate have been monitored through IoT in Abrasive Water Jet Machining (AWJM) processes. The different sensors, micro controller, wireless sensor network and modules have been used to improve the machinability of Haynes alloy. The machining performances have been evaluated with and without IoT system. Hence, the total process is called as smart machinability system which is applied on AWJM process. © 2021 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on Advances in Materials Research - 2019.

World University root node

Chennai Campus

Materials Today: Proceedings

IoT based smart machinability system and performances of MRR on Haynes alloy

In VLSI system design power consumption and energy dissipation are become more important. To optimize the power consumption we are using adiabatic logic which is having capability of reusing power. In this paper we proposed novel 7T SRAM cell using MCPL adiabatic logic. We compared the performance of energy dissipation and power consumption of conventional 7T SRAM and adiabatic SRAM. Furthermore, the conventional 6T,7T SRAM cell and the designs which are related are designed using FINFET devices to achieve low power.7T SRAM is more stable compared to 6T SRAM but due to excess transistor in 7T SRAM the power consumption will be more. SRAM cells are designed in cadence EDA environment and layouts for these circuits are carried out in cadence Assura tool. © 2021 Institute of Physics Publishing. All rights reserved.

Fulltext

Journal of Physics: Conference Series

Design of low power SRAM cell using adiabatic logic

In humans, the abnormality in brain arises due to various reasons and the ischemic-stroke (IS) is one of the major brain syndromes to be diagnosed and treated with appropriate procedures. The brain-signals and brain-images are widely considered for the clinical level diagnosis of IS. The proposed research considered the brain-image (MRI) based assessment of IS, due to its accuracy and multi modality nature. The MRI slices with modalities, such as diffusion-weighted (DW), flair and T1 are considered for the assessment. This work implements the following procedures to extract the IS lesion (ISL); (i) pixel level image fusion based on principal-component-analysis (PCA), (ii) image thresholding using cuckoo-search (CS) and Tsallis entropy, (iii) watershed based ISL extraction, and (iv) comparison of segmented ISL with the ground-truth-image (GTI). To confirm the clinical significance of the proposed work, the test images are collected from the benchmark ISLES2015 database. The results of this research confirms that, the fused brain MRI slices with DW and flair (DW + flair) modality facilitate to attain improved mean values of Jaccard-Index (83.17 ± 7.32%), Dice (88.51 ± 4.76%) and segmentation accuracy (97.34 ± 1.62%) compared to other images. This research confirms that, pixel level fusion will help to achieve better result during the clinical level disease diagnosis. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH, DE part of Springer Nature.

Evolutionary Intelligence

Image fusion practice to improve the ischemic-stroke-lesion detection for efficient clinical decision making

Semantic description of scenes is a challenging problem in video retrieval. Early works enlarged the semantic gap owing to the retrieval mechanism based on text matching between textual metadata provided by the up loader and text-based query. The need for bridging the semantic gap between user’s query content and retrieved videos prompted the research in content-based semantic video retrieval. This chapter proposes a semantic video retrieval model for human habitual scene search without textual metadata annotation. The proposed semantic model relies on understanding of the query content in the video to retrieve user’s query over a highly challenging video database. This work encompasses automatic concept/object detection in frames using concept bank, relates the detected concepts using concept occurrence vector to retrieve the relevant videos for the user’s query. The proposed semantic model relies on pedagogical understanding of the query content in the video using the prototype to retrieve users query over a highly challenging video database by poisson binomial radius measure. To evaluate the proposed approach, we validated the concept bank driven semantic retrieval model on user-created database for the user’s query by example. The experimental observation shows the astounding performance of the proposed concept knowledge-based semantic retrieval model. © 2021, Springer Nature Singapore Pte Ltd.

Lecture Notes in Electrical Engineering

Concept Bank Driven Semantic Video Retrieval Model

Accurate state of charge (SOC) estimation is considered as the main barrier to adopt Lithium-ion battery-based electric vehicles as alternatives to conventional internal combustion engine vehicles. There are several SOC estimation techniques recommended by many researchers still; they are all not accurate. This paper presents a robust and accurate SOC estimation battery model developed using an independently recurrent neural network (IndRNN). The proposed SOC estimation battery model is developed by training IndRNN constructively under various experimental datasets collected from Lithium Nickel Cobalt Aluminium Oxide battery cell at different ambient temperatures. Without any prior knowledge about battery internals, the proposed battery model successfully characterizes the non-linear behaviour of the battery effectively. Furthermore, its performance is proved by comparing with similar RNN architecture such as gated recurrent unit (GRU) and long short-term memory (LSTM). The attained results demonstrate that an IndRNN outperformed both GRU and LSTM in terms of accuracy under different electric vehicle drive cycle with minimal root mean square error of 0.7633% and mean absolute error of 0.6389% for diverse temperature conditions. Abbreviations: ANN: artificial neural network; BMS: battery management system; BPNN: back-propagation neural network; DFS: deep feature selection; EV: electric vehicle; FL: fuzzy logic; FNN: feedforward neural network; GA: genetic algorithm; GRU: gated recurrent unit; HWFET: highway fuel economy test; ICE: internal combustion engine; IndRNN: independently recurrent neural network; LA92: Los Angeles 92; LiB: lithium-ion battery; LSTM: long short-term memory; MAE: mean absolute error; MAX: maximum error; RNN: recurrent neural network; RMSE: root mean square error; ReLU: rectified linear unit; SOC: state of charge; SOH: state of health; SVM: support vector machine; US06: supplemental federal test procedure; UDDS: urban dynamometer driving schedule. © 2021 IETE.

Department of Embedded Technology

Vellore Campus

IETE Journal of Research

State of Charge Estimation of Lithium Batteries in Electric Vehicles Using IndRNN

In this article, the performance of the multiple-input multiple-output (MIMO) based free space optical (FSO) link is analyzed with Gamma-Gamma (GG) distribution in terms of Bit Error Rate (BER). The solution for the GG distribution with MIMO leads to complexity as it involves Meijer-G function. To reduce the complexity, we propose a mathematically tractable power series based probability density function (PDF) of the GG fading with Pointing Errors (PEs) and Atmospheric Turbulences (ATs). With the proposed PDF the exact BER is derived and the performance of the MIMO link with M-ary Phase Shift Keying (M-PSK) is analyzed in terms of BER and SNR. The maximal ratio combining (MRC) provides high SNR gain of 3. 5dB over equal gain combining (EGC) and 8dB SNR gain over selection combining (SC) for negligible PEs in=11.23 at BER 10-6. The proposed power series based PDF offers high SNR gain and power efficiency of the MIMO/FSO link over SISO and SIMO link. © 2021 IEEE.

2021 International Conference on COMmunication Systems and NETworkS, COMSNETS 2021

Performance Analysis of Power Series based MIMO/FSO Link with Pointing Errors and Atmospheric Turbulence

COVID-19 is a deadly viral infection that has brought a significant threat to human lives. Automatic diagnosis of COVID-19 from medical imaging enables precise medication, helps to control community outbreak, and reinforces coronavirus testing methods in place. While there exist several challenges in manually inferring traces of this viral infection from X-ray, Convolutional Neural Network (CNN) can mine data patterns that capture subtle distinctions between infected and normal X-rays. To enable automated learning of such latent features, a custom CNN architecture has been proposed in this research. It learns unique convolutional filter patterns for each kind of pneumonia. This is achieved by restricting certain filters in a convolutional layer to maximally respond only to a particular class of pneumonia/COVID-19. The CNN architecture integrates different convolution types to aid better context for learning robust features and strengthen gradient flow between layers. The proposed work also visualizes regions of saliency on the X-ray that have had the most influence on CNN's prediction outcome. To the best of our knowledge, this is the first attempt in deep learning to learn custom filters within a single convolutional layer for identifying specific pneumonia classes. Experimental results demonstrate that the proposed work has significant potential in augmenting current testing methods for COVID-19. It achieves an F1-score of 97.20% and an accuracy of 99.80% on the COVID-19 X-ray set. © 2020 Elsevier B.V.

Applied Soft Computing

Learning distinctive filters for COVID-19 detection from chest X-ray using shuffled residual CNN

Precise delineation of the ischemic lesion from unimodal Magnetic Resonance Imaging (MRI) is a challenging task due to the subtle intensity difference between the lesion and normal tissues. Hence, multispectral MRI modalities are used for characterizing the properties of brain tissues. Traditional lesion detection methods rely on extracting significant hand-engineered features to differentiate normal and abnormal brain tissues. But the identification of those discriminating features is quite complex, as the degree of differentiation varies according to each modality. This can be addressed well by Convolutional Neural Networks (CNN) which supports automatic feature extraction. It is capable of learning the global features from images effectively for image classification. But it loses the context of local information among the pixels that need to be retained for segmentation. Also, it must provide more emphasis on the features of the lesion region for precise reconstruction. The major contribution of this work is the integration of attention mechanism with a Fully Convolutional Network (FCN) to segment ischemic lesion. This attention model is applied to learn and concentrate only on salient features of the lesion region by suppressing the details of other regions. Hence the proposed FCN with attention mechanism was able to segment ischemic lesion of varying size and shape. To study the effectiveness of attention mechanism, various experiments were carried out on ISLES 2015 dataset and a mean dice coefficient of 0.7535 was obtained. Experimental results indicate that there is an improvement of 5% compared to the existing works. © 2020, Korean Society of Medical and Biological Engineering.

Computer Science

School of Computer Science and Engineering

Biomedical Engineering Letters

Delineation of ischemic lesion from brain MRI using attention gated fully convolutional network

One of the most important hardware blocks in processors is the multiplier circuit module. Area, power, and delay rule as primary factors deciding VLSI design methodologies. The work presented in this paper focuses on these three factors using Vedic multiplication approach and gate-diffusion input (GDI) logic. The ancient system of Indian mathematics being the Vedic mathematics, rediscovered from the Vedas, is more simplified, faster and accurate as compared to normal multiplication methods. The primary advantage of gate-diffusion input (GDI) logic is helping in reduced transistor count. Hence, the combination of these approaches of Vedic multiplication implemented using GDI logic results in reduced propagation delay time, lower power consumption, and less silicon area. In this paper, the proposed 2-bit Urdhva cell used is implemented using AND gate, XOR gate, and an inverter unlike the existing literature in which it is implemented using two half adders. The results are validated for the comparative advantages of our approach. The circuit simulations are carried out using UMC 90 nm technology nodes in Cadence Virtuoso. © 2021, Springer Nature Singapore Pte Ltd.

4-Bit Vedic Multiplier Design Using Gate-Diffusion Input (GDI) Logic

A compressive sensing (CS) approach is proposed for the complex wave retrieval algorithm from single off-axis digital Fresnel holography to improve the accuracy of quantitative phase measurement. The linear model of non-linear holographic process used in the complex wave retrieval algorithm is utilized to meet the linearity requirement of the proposed CS implementation presented in this paper. The error in object wave reconstruction due to the approximations used in complex wave retrieval algorithm is compensated by combining it with CS approach. A Mach–Zehnder interferometric geometry in transmission and off-axis mode in the Fresnel domain is used to record the digital hologram. The numerical simulations and experimental results are presented to demonstrate and validate the proof of the concept and to compare CS frame work with conventional complex wave retrieval algorithm. The proposed algorithm is tested for the microlens sample characterization by considering the parameters such as height, diameter and radius of curvature. A Haar wavelet based object wave sparsification is used in the experiment for the prudent phase reconstruction in the CS approach. The proposed method enables an improved and accurate object phase reconstruction, with minimal noise such as distortions or speckles with respect to the conventional complex wave retrieval method. © 2020 Elsevier B.V.