Showing papers by "Motilal Nehru National Institute of Technology Allahabad published in 2023"

Machine Learning Approaches for the Detection of Schizophrenia Using Structural MRI

Abstract: The reproducibility of Computer Aided Diagnosis (CAD) in detecting schizophrenia using neuroimaging modalities can provide early diagnosis of the disease. Schizophrenia is a psychiatric disorder that can lead to structural abnormalities in the brain, causing delusions and hallucinations. Neuroimaging modality such as a structural Magnetic Resonance Imaging (sMRI) technique can capture these structural abnormalities in the brain. Utilizing Machine Learning (ML) as a potential diagnostic tool in detecting classification biomarkers can aid clinical measures and cater to recognizing the factors underlying schizophrenia. This paper proposes an ML based model for the detection of schizophrenia on the structural MRI dataset of 146 subjects. We sought to classify schizophrenia and healthy control using five ML classifiers: Support Vector Machine, Logistic Regression, Decision Tree, k-Nearest Neighbor, and Random Forest. The raw structural MRI scans have been pre-processed using techniques such as image selection, image conversion, gray scaling of MRI images, and image flattening. Further, we have tested the performance of the model using hold-out cross-validation and stratified 10-fold cross-validation techniques. The results showed that the SVM achieved high accuracy when the dataset was validated using a stratified 10-fold cross-validation technique. On the other hand, k-Nearest Neighbor performed better when the hold-out validation method was used to evaluate the classifier.

Ecological Footprint Assessment of e-Waste Recycling

Abstract: Rapid technological advancements and socio-economic developments have given rise to the ever-increasing demands of electronic devices. These devices produce a huge amount of E-waste at the end of their operational life. According to the United Nations Global E-waste Monitor 2020 report, the annual generation of e-waste has grown up by 21% in the last five years. Recycling of e-waste is a sustainable measure to protect the ecosystem from its hazardous effects, but the recycling itself has a significant ecological impact on the planet. Prompt growth of e-waste recycling has created an urgent need for quantitative environmental assessment of e-waste recycling. The study proposed a methodology to assess the Ecological Footprint (EF) of e-waste recycling. The total EFER of E-waste recycling is estimated as 0.023 gha/tonne. The carbon absorption land contributes the highest among all other bio-productive lands for the e-waste recycling. Production and use of electronic devices will increase in near future and the e-waste recycling impact will also be scaled accordingly.

Herbal Therapeutics for Alzheimer's Disease: Ancient Indian Medicine System from the Modern Viewpoint

Abstract: Abstract: Alzheimer's is a chronic neurodegenerative disease where amyloid-beta (Aβ) plaques and neurofibrillary tangles are formed inside the brain. It is also characterized by progressive memory loss, depression, neuroinflammation, and derangement of other neurotransmitters. Due to its complex etiopathology, current drugs have failed to completely cure the disease. Natural compounds have been investigated as an alternative therapy for their ability to treat Alzheimer's disease (AD). Traditional herbs and formulations which are used in the Indian ayurvedic system are rich sources of antioxidant, anti-amyloidogenic, neuroprotective, and anti-inflammatory compounds. They promote quality of life by improving cognitive memory and rejuvenating brain functioning through neurogenesis. A rich knowledge base of traditional herbal plants (Turmeric, Gingko, Ashwagandha, Shankhpushpi, Giloy, Gotu kola, Garlic, Tulsi, Ginger, and Cinnamon) combined with modern science could suggest new functional leads for Alzheimer's drug discovery. In this article Ayurveda, the ancient Indian herbal medicine system based on multiple clinical and experimental, evidence have been reviewed for treating AD and improving brain functioning. This article presents a modern perspective on the herbs available in the ancient Indian medicine system as well as their possible mechanisms of action for AD treatment. The main objective of this research is to provide a systematic review of herbal drugs that are easily accessible and effective for the treatment of AD.

Prediction of SOH and RUL for Lithium-Ion Batteries Using Regression Method with Feature of Indirect Related to SOH (FIRSOH) and Linear Time Series Model

Abstract: Lithium-ion battery state of health (SOH) and residual usable life (RUL) are two significant parameters that are typically estimated using battery capacity. The capacity of lithium-ion batteries for online applications can be measured indirectly. In this study, voltage, current, and temperature curves during charging and discharging of lithium-ion batteries—which react to the process of battery capacity degradation—are used to extract Feature of indirect related to SOH (FIRSOH). The Linear Ridge Regression (LRR) method is used for SOH prediction. Then, taking into account that SOH and RUL have a particular mapping relationship, five FIRSOHs and the current SOH value are used to forecast RUL of lithium-ion batteries using the linear time series (LTS) model. The outcomes demonstrate that the suggested approach has a high level of prediction accuracy.

Design of Low-Power High Performance SAR Analog-To-Digital Converter Based on MOS Capacitor DAC Structure

Abstract: In this article, a 5-bit 20 MS/s successive approximation register (SAR) analog-to-digital converter (ADC) with 4.53 bits ENOB and 29.0 dB SNDR is designed using 90 nm CMOS process technology using CADENCE virtuoso. This SAR ADC consumes power of 31.67 µW with the figure-of-merit (FoM) 68.54 fJ/conv-step. A modified dynamic comparator with two inverters connected at the output to avoid metastable state of sampled signal is used, which also helps in reducing power consumption. A binary-weighted charge redistribution digital-to-analog converter (DAC) is used using metal oxide semiconductor (MOS) capacitor to lessen the power dissipation also less area requirement as compared to conventional metal–insulator-metal (MIM) capacitor-based DAC. This SAR ADC has potential application for low power and low to medium sampling rate biomedical devices.

Blockchain-based Secure Privacy-Preserving Vehicle Accident and Insurance Registration

Abstract: Insurance claims processing involves multiple entities and data sources, necessitating communication between human agents. Consequently, vehicle insurance claims have traditionally required significant human effort and time. Daily vehicle-related transactions, including those managed by transportation authorities, pose challenges for tracking. Centralised systems have been utilised for national solutions, but trust management, transparency, and access control issues arise. There is potential for further integration of vehicle-related transactions. This article proposes a blockchain framework for vehicle insurance to streamline the reporting of accidents and filing of insurance claims. Blockchain-based automation platforms can enhance the scale and response time of claims processing, providing users with control over additional transactions, inspection, and insurance. For experimental purposes, a blockchain was created using Hyperledger Fabric to store information about vehicles, owners, and insurance. Efficient querying of this blockchain requires specific participants, assets, and transactions. The consensus algorithm can identify invalid claims if a transaction request contains an error. By deploying blockchain technology and smart contracts, this architecture has the potential to address trust and security concerns associated with traditional insurance policies and claims.

Design of SRAM Cell for Improved Performance

Abstract: In this document, a low power 13 Transistor Static Random Access Memory (SRAM) has been proposed to improve write access time as compared to that from the reference 11T SRAM Cell which offers single-ended write operation, by adding a PMOS and NMOS transistor on BL side. Simulation results in Cadence Virtuoso tool using 90 nm GPDK Technology, shows that the proposed cell has reduced write access time by 39%. Also, its showing reduced DC power dissipation by 5.9% and read power dissipation by 70%.

Low Power Radix-4 Booth Multiplier Design Using Pass Transistor Logic

Abstract: The multiplier is the core part of the processor and embedded systems. Multiplication is intensive on hardware, and the criteria of interest are higher performance and low-power arithmetic structure of the processor. If the power consumed by the multipliers is reduced, the overall power consumption in the processor is reduced drastically. The projected multiplier circuit comprises a partial product generator (PPG) and adder trees. The design of PPG and adder tree can be optimized separately. This work proposes a novel Radix-4 Booth Multiplier PPG technique using Pass Transistor Logic. The method using PTL helps reduce the transistor count and the delay in the circuit. The design uses a pre-encoding phase which switches off the encoder in zero state to save power. The gates are modified to include CMOS inverters in intermediate design parts, which helps level the circuit and reduce the noise from the input. The parameters compared are the delay in generating the partial product power consumption by the partial product generator and the power delay product of the output signal.

Design and Implementation of Matched Filtering and Timing Recovery Algorithm for IEEE 802.15.4 Digital Baseband Front End

Abstract: IEEE 802.15.4 is a technical standard for low-rate wireless personal area network for long range communication. This standard provides a framework for the lower layers (PHY and MAC) in the OSI model for wireless connectivity network which is used in IoT applications. In a communication network, channel noise and timing mismatch are common problems. In order to recover the received signal successfully, a receiver should be able to overcome these problems. Demodulation and demapping alone would not be able to serve the purpose. This paper proposes modeling, design, and simulation of signal conditioning blocks in the receiver for O-QPSK PHY and SUN O-QPSK PHY mentioned in the standard. These blocks process the signal available at the digital baseband front end of the receiver which is available after RF to analog baseband down conversion and analog baseband to digital baseband conversion. Signal conditioning introduced in this work involves the design of matched filter and timing recovery to overcome the noise and sampling time mismatch. These blocks along with the O-QPSK transceiver have been first modeled in MATLAB/Simulink. Logic design for the modules introduced in the receiver chain has been implemented using Verilog RTL and verified through simulation. For FPGA implementation, RTL modules of matched filter and timing recovery have been integrated at the output of ADC, and output of timing recovery goes for inverse pulse shaping and is fed to O-QPSK and smart utility network (SUN) O-QPSK demodulator of receiver chain. All the modules are implemented on the FPGA Board Xilinx UltraScale + VCU118.

Review of the SAR ADC’s Energy Efficient Switching Scheme

Abstract: This article presents an overview on different switching methods for improvement of energy and area efficiency of Successive-Approximation-Register type Analog-to-Digital Converters (SAR ADCs). Firstly, conventional switching method is discussed then all other methods of switching are explained and compared with reference of conventional switching. In comparison to conventional switching, various switching procedures are compared for logic complexity, energy efficiency, and area savings. Monotonic and tri-level schemes are most energy efficient among all other switching methods but at the cost of degradation in linearity of SAR ADC. For linear SAR ADC applications merged capacitor switching (MCS) and detect-and-skip switching method can be used because it has no common-mode voltage variation. This paper’s primary goal is to give designers a brief introduction of various switching techniques so they can use them to reduce switching energy consumption in Digital to Analog Converter (DAC) arrays in SAR ADCs.

Minimal Group Delay Multi-objective Finite Impulse Response Filter Design Using Salp Swarm Algorithm and Its Improved Version

Abstract: A recent trend suggests that assessing low complexity is critical for effective digital filter design. Previously, despite the requirement for low implementation complexity in FIR filter design, current techniques only focused on decreasing pass band and stop band ripples. In this situation, a multi-objective improved salp swarm optimization technique has been used to create an optimization-based FIR filter. The proposed solution not only reduces pass band and stop band ripples, but it also reduces of group delay the number of multipliers required to create a FIR filter that adheres to the desired frequency range. The Pareto front is a set of optimum solutions in the non-dominated solution of filter design that keeps the trade-off amid the several criteria. To validate the pertinency of the suggested process, the constructed FIR filter was compared to other documented methods. The evaluation and analysis were carried out in a regrading manner.

IoT for Landslides: Applications, Technologies and Challenges

Abstract: The most common existing human-induced and natural geohazard causing damage to infrastructure and property loss are landslides occurring in hilly regions of India. Landslide involves excessive surface movements, including rock failure, slope failure, debris flow, etc. Landslide prevention is based on the principles of slope stability engineering and developing techniques to reduce its effects on natural resources, river ecosystems, and infrastructure loss. Early warning and monitoring of landslides are the most critical detection/prevention strategies. With advancements in the evolution of the Internet of Things (IoT) technique, the recent innovative concept involves the utilization of IoT to strengthen particularly capability and accuracy of early warning and monitoring frameworks to prevent landslides. This chapter presents a brief overview of related research and technological advancements in IoT techniques used for landslide studies. First, it studies the importance and application of IoT for landslide monitoring and early warning system design, then investigates the fundamental layers in IoT along with the application and key technologies involved in landslide prevention. This chapter also highlights challenges with IoT-based landslide early warning and monitoring systems.

Development of an Automated Algorithm to Quantify Optic Nerve Diameter Using Ultrasound Measures: Implications for Optic Neuropathies

Abstract: The paper presents a novel computational and image processing algorithm for automatic measurement of optic nerve diameter (OND) from B-scan ultrasound images acquired in a traumatic cohort. The OND is an important diagnostic parameter for the detection and therapeutic planning of several diseases and trauma cases. The automated measurement of OND may provide reliable information for predicting intracranial pressure (ICP) in traumatic patients. In the proposed method, the automatic measurement of the OND involves pre-processing an ultrasound image, followed by retinal globe detection, optic nerve localization, optic nerve tip detection and measurement of OND. The proposed algorithm measures the OND in the perpendicular direction to the optic nerve axis, considering the orientation of the optic nerve in an automated framework. The developed algorithm automatically calculates the OND. The study includes Twenty-four traumatic individuals with optic nerve pathologies whose B-scan ultrasound images of the optic nerve were manually obtained in the axial plane. The accuracy of the automatic measurement has been quantified by comparing it with manually measured OND by medical experts. A low Percent Root Mean Square Difference (PRD) value of 14.92% between automated and manually measured OND is found as an accuracy measure for the automatic measurement of OND. The proposed algorithm automatically and accurately determines the optic nerve diameter from an eye ultrasound image.

Selective Harmonic Compensation in Active Power Filter Using Nonlinear Predictive Current Control Method

Abstract: In this chapter, a non-linear predictive current control (PCC) scheme is proposed based on the inner current loop and outer voltage loop using selective harmonic compensation for single-phase shunt active power filters (SAPFs). This study designs the controller parameters utilizing conventional non-linear techniques, namely high-frequency rejection transfer functions (combination of band-pass filter and harmonic compensator transfer function). The proposed control algorithm is executed and a comparative study between the conventional and proposed technique in APFs has been presented. The novelty of the proposed PCC technique is the selective harmonic elimination from the grid current which comes from non-linear loads. In this chapter, the third harmonic component is eliminated from the grid current. The proposed technique is simulated on a virtual HIL environment and real-time simulation validation is done by using FPGA based Typhoon HIL 402 kit.

Robust and Imperceptible Color Image Watermarking Using LWT, Schur Decomposition, and SVD in YCbCr Color Space

Abstract: A robust and imperceptible color image watermarking scheme is presented to conceal the data by taking the QR code logo as the watermark. The YCbCr color space is used for watermarking. The proposed method uses the fundus image as the input host image. The fundus image is the image of the rear part of the eye. Fundus photography is done with the help of specialized cameras. The suggested method for adding a watermark uses the Lifting Wavelet Transform (LWT), Schur Decomposition, and Singular Value Decomposition (SVD). Even if the recovered QR code image is deformed, it is a resilient code from which contained data can be recovered. The proposed watermarking technique is robust against attacks like JPEG Compression, Region of interest filtering, Salt and Pepper noise, Average filter, Gaussian noise, and Median filter. To check the performance of the proposed technique, Peak signal-to-noise ratio (PSNR), Mean square error (MSE), Normalized correlation coefficient (NCC), and Structural similarity index measurement is used.

Study and Analysis of Three-Stage Single-Miller CMOS OTA

Abstract: This paper describes a three-stage CMOS operational transconductance amplifier (OTA), high-voltage (HV) operational amplifier, active single-Miller capacitor, and inner half feed-forward (ASMIHF) amplifier. Frequency compensation network employs a single-Miller capacitor, and an inverting current buffer is incorporated in the input stage of the amplifier. A feed-forward route and a slew rate enhancement circuit are used to extend the large-signal quick responsiveness. The suggested OTA does not have a derivable capacitive load (CL) upper limit, in line with an in-depth small-signal study. The OTA is formed using conventional 90-nm technology and encompasses a die area of 0.0027 mm2. These OTA circuits are analyzed and compared in terms of parameters like DC gain, unity-gain frequency, phase margin, gain margin, and slew rate.

Design of Low Power High Speed 2nd Order Discrete Time Sigma-Delta Modulator Using Charge Shared Double Tail Dynamic Comparator

Abstract: In this paper, we study a switch capacitive second-order passive-active sigma-delta modulator with twin benefits of improved speed and lower power dissipation. The comparator block has been on prime focus to improve power and speed. Two design variants of integrators are utilized for improvement. As the filter is simply switch capacitive integrator, there will be high attenuation with lower power dissipation. The second stage integrator is followed by unity gain buffer which does compensation of gain errors and phase errors with respect to ideal integrator transfer function. A 2nd order discrete time sigma-delta modulator will yield exemplary results with using large oversampling ratio. This circuit contains double tail dynamic comparator based on charge sharing scheme to reduce power dissipation and improving speed for the modulator. For circuit level verification, the circuit is simulated using Cadence Virtuoso using GPDK 90 nm CMOS technology. The measurements depicts power dissipation of 1.44 µW, peak Signal to noise ratio & signal to noise and distortion ratio of 84.67 dB and 82.95 Db, respectively, dynamic range is 91.02 dB & ENOB is 13.48 bits using 1 V supply for a 500 Hz sinusoidal signal.

Enhanced Energy-Aware Fault Tolerance Technique for Real-Time Task on Heterogeneous Multicore System

Abstract: Real-time system is used in various safety-critical areas where any single fault can damage the whole system. So, fault tolerance must be taken care of. Fault tolerance is a way to make system to continue the work even in the presence of fault or error and complete the task in predetermined time. These fault can be hardware or software fault. Several fault tolerance technique have been proposed. But, in the process of fault tolerance, extra resources are used which increase the energy consumption of system. So, these two terms energy management and fault tolerance both are orthogonal to each other. Also, for real-time system, both are very critical. Here, we are going to cover energy-aware and fault tolerant scheduling of periodic task in a heterogeneous multicore system. For fault tolerant, enhanced MPB-PS algorithm is proposed with speed fine-tuning test. To reduce the energy consumption, DVFS is applied to scale the primary tasks and increases the chances to cancel the backup task. A comparative analysis of various method is discussed and proposed a scheme which satisfy the real-time constraints as well as fault tolerance requirements of real-time tasks in a more energy-efficient way.

Abnormality Detection in Heart Using Combination of CNN, RNN and U-Net

Abstract: Nowadays when people are rapidly progressing in technology and various other departments, they are forgetting very basic thing, i.e., living standards and health. World is facing malnutrition and illness at very large scale. Heart disease is one of them where issues are increasing day by day. Due to very high pressure of competitive environment, people are having cardiac arrest at normal basis. Therefore, proposed work develops a model which can predict whether a person has heart disease or not by using CMR images. This model is helpful for everyone to diagnose their disease even at home, so that they can be treated immediately. If someone loads faded images or cropped images, the proposed model can even predict the disease without any limitations.

Ecological Footprint Assessment of Concrete Using e-Waste

Abstract: Rapid technological enhancement has become a cause of e-waste generation and will also create problems for the environment in near future. In this study, the environmental assessment of concrete using e-waste has been examined. The Ecological Footprint of the M20 grade plain cement concrete is calculated as 0.04678 gha/m3, however, if 5–25% of aggregate (by volume) is replaced by e-waste in the concrete, the Ecological Footprint of the recycled plain cement concrete reduces by 0.12–0.59%. The partial replacement of the aggregate by e-waste in plain cement concrete reduces the concrete weight by 2.1–10.1%. The Sustainable Recycling Index (SRI) is also developed for the environmental assessment of e-waste as a replacement for aggregate in plain cement concrete. The SRI value of 5%, 10%, 15%, 20% and 25% aggregate replacement with e-waste are 1.6 × 10–5, 6.4 × 10–4, 1.4 × 10–4, 2.6 × 10–4 and 4 × 10–4; respectively. The production of concrete using e-waste provides a sustainable option for e-waste assimilation. Therefore, concrete using e-waste could be seen as an eco-friendly, lightweight, and economical alternative to the conventional concrete.

Improved Fuzzy Logic-Based Localization in Wireless Sensor Networks

Abstract: One of the very fundamental properties of a sensor node is its location. The location of the wireless sensor node is an important parameter to interact efficiently and get the most out of the wireless sensor network (WSN). Finding the location information of a sensor node is called the localization problem. In this study, a received signal strength (RSS)-based distance estimation approach is employed for processing signal strength measurements to approximate the node’s location. The distances from the anchor node are calculated using the fuzzy inference system (FIS), which takes RSS as input and gives the corresponding distance as output. The trilateration method is employed to calculate the estimated coordinates of sensor nodes. We compare methods that employ Mamdani and Sugeno fuzzy inference systems with triangular and trapezium membership functions to calculate the distance from RSS values and simulate the results to demonstrate the respective average localization error.

Oncology: Way Forward from Clinical and Molecular Diagnosis to Treatment

Abstract: Traditional methods of treating cancer, such as surgery, chemotherapy, and radiation, are now frequently used in combination with one another as well as medications that target certain biological networks. The area of oncology has changed dramatically over the past 10 years as a result of developments in our understanding of the cancer biology. Clinical oncology uses molecular assays in a variety of ways. Hereditary cancer disorders are now often diagnosed via mutation analysis. Strict medical supervision and a variety of preventative measures are beneficial for healthy people who carry cancer-predisposing mutations. The approach to treating cancers brought on by germ-line alterations must be significantly changed. On the basis of the existence of identifiable mutations, cancer therapy now includes the personalised selection of cancer medications. Genetic examination is no longer restricted to obvious chromosomal defects in karyotypes, DNA in cells may now be examined down to the level of individual base pairs. The likelihood that individualised treatment for certain malignancies may be available soon rises as a result of our deep understanding of cancer genetics. In this chapter, the nucleic, cytogenetic, and bioinformatic methods that are expected to have an impact on cancer treatment and management are being discussed.

A Non-invasive Planar Resonant Microwave Sensor for Unknown Liquid Permittivity Estimation

Abstract: This paper presents a split ring-based resonant microwave sensor for precise measurement of liquid sample’s permittivity. The modified SRR sensor is etched on the substrate top plane and is excited by electromagnetic coupling by the microstrip line on the same plane. The liquid sample having unknown permittivity value when placed over sensor alters the field distribution near the sensor region which shifts the sensor’s natural resonant frequency. The correlation between frequency shift and the permittivity of the liquid sample has been used to estimate permittivity of unknown sample. Further, a closed form numerical expression has been developed using curve fitting tool. Validation of the developed numerical model has been successfully done utilizing different independent liquid samples, and a slight error (less than 6%) was reported between actual and calculated permittivity values. The developed sensor is quite simple, compact, cheap, easy to fabricate and highly sensitive, covering the wide permittivity range of liquids for permittivity estimation.

An Imperceptible Semi-blind Color Image Watermarking Using RDWT and SVD

Abstract: A semi-blind color image watermarking scheme using redundant discrete wavelet transform (RDWT) and singular value decomposition (SVD) is proposed in this paper. Red, green, and blue channels are extracted from the colored cover image, and 1-level RDWT is applied to decompose it into four sub-bands. LL sub-band is used to embed color watermark by modifying their singular value. At the extraction end, the reverse process of embedding algorithm is followed. Various test images and fundus images are used to evaluate the performance of the proposed technique. Performance parameters peak signal–to-noise ratio (PSNR), structural similarity (SSIM), and normalized correlation coefficient (NC) are measured. Simulation results show the imperceptibility and robustness analysis. Results are also compared with existing techniques, and the proposed method outperforms these techniques.

Performance Analysis of mmWave V2V Communication Using Relay Vehicle for Advanced Safety Applications

Abstract: Advanced V2X communication applications require very stringent communication performance parameter conditions like data rate, end-to-end delay, data dropping ratio and reliability of the communication link to fulfil the requirements of the respective advanced applications as depicted by their corresponding 3GPP Use Case Group. The current IEEE 802.11p and 3GPP LTE C-V2X are unable to meet these stringent conditions. Communication at mmWave frequency ranges has the capability of meeting these advanced application requirements, but suffers from high path and penetration loss of the signal. Therefore, to increase the QoS of the V2V communication, the use of intermediate vehicle between source vehicle and destination vehicle becomes vital. Analytical Hierarchical Process (AHP) and Coalition Game (CG) theory are used for the selection of relay vehicles from a set of candidate relay vehicles. These relay vehicles (or vehicular User Equipment) provide the path for message transfer when LOS path is blocked between source vehicle and destination vehicle or when the LOS distance between them is too much to maintain the QoS requirements of the particular advanced V2X application.