Showing papers by "College of Engineering, Pune published in 2020"

Forecasting Models for Coronavirus Disease (COVID-19): A Survey of the State-of-the-Art

Abstract: COVID-19 is a pandemic that has affected over 170 countries around the world. The number of infected and deceased patients has been increasing at an alarming rate in almost all the affected nations. Forecasting techniques can be inculcated thereby assisting in designing better strategies and in taking productive decisions. These techniques assess the situations of the past thereby enabling better predictions about the situation to occur in the future. These predictions might help to prepare against possible threats and consequences. Forecasting techniques play a very important role in yielding accurate predictions. This study categorizes forecasting techniques into two types, namely, stochastic theory mathematical models and data science/machine learning techniques. Data collected from various platforms also play a vital role in forecasting. In this study, two categories of datasets have been discussed, i.e., big data accessed from World Health Organization/National databases and data from a social media communication. Forecasting of a pandemic can be done based on various parameters such as the impact of environmental factors, incubation period, the impact of quarantine, age, gender and many more. These techniques and parameters used for forecasting are extensively studied in this work. However, forecasting techniques come with their own set of challenges (technical and generic). This study discusses these challenges and also provides a set of recommendations for the people who are currently fighting the global COVID-19 pandemic.

Artificial Intelligence and Machine Learning in Computational Nanotoxicology: Unlocking and Empowering Nanomedicine.

Abstract: Advances in nanomedicine, coupled with novel methods of creating advanced materials at the nanoscale, have opened new perspectives for the development of healthcare and medical products Special attention must be paid toward safe design approaches for nanomaterial-based products Recently, artificial intelligence (AI) and machine learning (ML) gifted the computational tool for enhancing and improving the simulation and modeling process for nanotoxicology and nanotherapeutics In particular, the correlation of in vitro generated pharmacokinetics and pharmacodynamics to in vivo application scenarios is an important step toward the development of safe nanomedicinal products This review portrays how in vitro and in vivo datasets are used in in silico models to unlock and empower nanomedicine Physiologically based pharmacokinetic (PBPK) modeling and absorption, distribution, metabolism, and excretion (ADME)-based in silico methods along with dosimetry models as a focus area for nanomedicine are mainly described The computational OMICS, colloidal particle determination, and algorithms to establish dosimetry for inhalation toxicology, and quantitative structure-activity relationships at nanoscale (nano-QSAR) are revisited The challenges and opportunities facing the blind spots in nanotoxicology in this computationally dominated era are highlighted as the future to accelerate nanomedicine clinical translation

Disturbance-Observer-Based Sensorless Control of PMSM Using Integral State Feedback Controller

Abstract: The objective of this article is to design robust speed control of a permanent magnet synchronous motor (PMSM). The mathematical model of the PMSM is highly nonlinear with uncertainties and disturbances. The most obvious method is to estimate these disturbances and design a robust controller to attenuate the effect of disturbances from the output. In this article, the load torque, which typically occurs as disturbance in the servo system, is estimated using a disturbance observer (DO). A nonlinear model of the PMSM is linearized using the Jacobian linearization around an operating point, and a state feedback controller with an integral term is designed at this operating point. Novelty of the proposed method lies in the way the DO is used to update this operating point, which is similar to the gain scheduling approach. To extend this method for a sensorless operation, a sliding-mode observer is used with the DO in cascade. The proposed scheme is validated using simulations on MATLAB/Simulink and an experimental setup using TMS320F28027. The results show improved performance as compared to traditional methods over a wide range of speed.

Artificial Intelligence and Machine Learning Empower Advanced Biomedical Material Design to Toxicity Prediction

Abstract: Materials at the nanoscale exhibit specific physicochemical interactions with their environment. Therefore, evaluating their toxic potential is a primary requirement for regulatory purposes and for the safer development of nanomedicines. In this review, to aid the understanding of nano–bio interactions from environmental and health and safety perspectives, the potential, reality, challenges, and future advances that artificial intelligence (AI) and machine learning (ML) present are described. Herein, AI and ML algorithms that assist in the reporting of the minimum information required for biomaterial characterization and aid in the development and establishment of standard operating procedures are focused. ML tools and ab initio simulations adopted to improve the reproducibility of data for robust quantitative comparisons and to facilitate in silico modeling and meta‐analyses leading to a substantial contribution to safe‐by‐design development in nanotoxicology/nanomedicine are mainly focused. In addition, future opportunities and challenges in the application of ML in nanoinformatics, which is particularly well‐suited for the clinical translation of nanotherapeutics, are highlighted. This comprehensive review is believed that it will promote an unprecedented involvement of AI research in improvements in the field of nanotoxicology and nanomedicine.

An efficient AES implementation using FPGA with enhanced security features

Abstract: Data transferred in an electronic way is vulnerable to attacks With an aim to protect data for secure communication, a new Hybrid non pipelined Advanced Encryption Standard (AES) algorithm based on traditional AES algorithm with enhanced security features is proposed in this work Abysmal analysis of the AES algorithm implies that the security of AES lies in the S-box operations This paper presents a new approach for generating S-box values (modified S-box) and initial key required for encryption/encryption (improved key generation) using PN Sequence Generator The AES algorithm with proposed modifications shows significant improvement in the encryption quality as compared to traditional AES algorithm The traditional AES algorithm equipped with proposed novel modified S-box technique and improved key generation technique gives an avalanche effect of 60% making it invulnerable to attacks The proposed design is synthesized on various Field Programmable Gate Array (FPGA) devices and compared to the existing designs resulting in significant improvement in throughput The proposed design is implemented on Spartan6 FPGA device

Deep Learning Based Representation for Face Recognition

Abstract: Face Recognition is one of the challenging process due to huge amount of wild datasets. Deep learning has been provided good solution in terms of recognition performance, as day by day this have been dominating the field of biometric. In this paper our goal is to study deep learning based face representation under several different conditions like lower and upper face occlusions, misalignment, different angles of head poses, changing illuminations, flawed facial feature localization using deep learning approaches. For extraction of face representation two different popular models of Deep learning based called Lightened CNN and VGG-Face and have reflected in this paper. As both of this model show that deep learning model is robust to different types of misalignment and can tolerate localizations error of the intraocular distance.

Revisited COVID-19 Mortality and Recovery Rates: Are we Missing Recovery Time Period?

Abstract: How difficult time it is due to COVID-19 pandemic can be determined by its mortality and/or recovery rates. Predictive modeling can help us forecast how big the impact will be, especially for human lives? Prediction is one of the wellknown studies that is entirely relying on machine learning based data analytics tools and techniques. In this study, we revisit mathematical models to measure severity levels of a COVID-19 pandemic. A pandemic is an epidemic occurring on a scale that spreads rapidly across the world. The World Health Organization (WHO) considers epidemic diseases are Chikungunya, Cholera, Crimean – Congo hemorrhagic fever, Ebola Virus disease, Hendra virus infection, Influenza, Lassa fever, Plague, COVID-19, SARS, etc. In 2014, the United States Centre for Disease Control and Prevention (CDC) announced an equivalent framework to the WHO’s pandemic stages titled pandemic intervals framework [1–3], where two pre-pandemic and four pandemic intervals were reported. Investigation and recognition are pre-pandemic intervals. Initiation, acceleration of diseases, deceleration, and preparation are pandemic intervals. In a similar fashion, instead of using pandemic intervals, we take recovery time period into account. Recovery time period can be two days, a week, two weeks, a month, six months, a year or any finite number of days. For COVID-19, we consider an average recovery time period of 14 days [4]. In March 2020, Baud et al. [4] reported to take 14 days delay into account in order to compute right mortality rate during the COVID-19 pandemic. Their study suggested that the classical mortality rate undervalue the probable threat due to COVID-19 in symptomatic cases. Of all metrics, to measure an austerity of the COVID-19 pandemic, in addition to mortality rates, recovery rate is considered. Both, mortality and recovery rates are useful metrics only when recovery time period is considered. Practically, one cannot compute mortality and recovery rates without considering its recovery time period. Inspired from previous work [4], we revisit mathematical models for both mortality and recovery rates, where not recovered cases and recovery time period are discussed thoroughly with both synthetic and real datasets. In the sequel, for any disease, mortality rate (MR) defines the probability of death, and can be expressed as, MR 1⁄4 N 100; where D and N refer to total number of deaths and infected people, respectively. While computing the MR, we consider the total number of an infected people till date, which will superfluously increase the denominator and hence decreases the MR. Therefore, practically, it is observed that the classical mathematical equation (to computeMR)may deviate from what it should be, since it does not take recovery time period into account. This primarily motivates us to revisit classical mathematical models. Since recovery time periods vary a lot from person to person (of course, demography dependent, as reported in WHO [5]), it is important to define an average recovery time period (Pavg) as follows: This article is part of the Topical Collection on Education & Training

Smart agriculture management system using internet of things

Abstract: In the world of digital era, an advance development with internet of things (IoT) were initiated, where devices communicate with each other and the process are automated and controlled with the help of internet. An IoT in an agriculture framework includes various benefits in managing and monitoring the crops. In this paper, an architectural framework is developed which integrates the internet of things (IoT) with the production of crops, different measures and methods are used to monitor crops using cloud computing. The approach provides real-time analysis of data collected from sensors placed in crops and produces result to farmer which is necessary for the monitoring the crop growth which reduces the time, energy of the farmer. The data collected from the fields are stored in the cloud and processed in order to facilitate automation by integrating IoT devices. The concept presented in the paper could increase the productivity of the crops by reducing wastage of resources utilized in the agriculture fields. The results of the experimentation carried out presents the details of temperature, soil moisture, humidity and water usage for the field and performs decision making analysis with the interaction of the farmer.

Frequency and pattern reconfigurable rectangular patch antenna using single PIN diode

Abstract: The paper presents the design of a frequency and pattern reconfigurable rectangular patch antenna using a single PIN diode switch. The use of single PIN diode reduces the complexity of the biasing network required for the diode. The proposed antenna can work in two different modes and resonates at 2.47 GHz, 3.8 GHz, and 5.36 GHz with the capability to change the radiation pattern. In this design biasing lines required for biasing of the diode are placed away from the radiating structure and the pattern search algorithm is used for the optimization of antenna feed for achieving impedance matching. These are the two improvisations made in the presented design as compared to the existing designs in literature.

Devanagari Handwritten Character Recognition using fine-tuned Deep Convolutional Neural Network on trivial dataset

Abstract: In order to rapidly build an automatic and precise system for image recognition and categorization, deep learning is a vital technology. Handwritten character classification also gaining more attention due to its major contribution in automation and specially to develop applications for helping visually impaired people. Here, the proposed work highlighting on fine-tuning approach and analysis of state-of-the-art Deep Convolutional Neural Network (DCNN) designed for Devanagari Handwritten characters classification. A new Devanagari handwritten characters dataset is generated which is publicly available. Datasets consist of total 5800 isolated images of 58 unique character classes: 12 vowels, 36 consonants and 10 numerals. In addition to this database, a two-stage VGG16 deep learning model is implemented to recognize those characters using two advanced adaptive gradient methods. A two-stage approach of deep learning is developed to enhance overall success of the proposed Devanagari Handwritten Character Recognition System (DHCRS). The first model achieves 94.84% testing accuracy with training loss of 0.18 on new dataset. Moreover, the second fine-tuned model requires very fewer trainable parameters and notably less training time to achieve state-of-the-art performance on a very small dataset. It achieves 96.55% testing accuracy with training loss of 0.12. We also tested the proposed model on four different benchmark datasets of isolated characters as well as digits of Indic scripts. For all the datasets tested, we achieved the promising results.

Data Analytics: COVID-19 Prediction Using Multimodal Data

Abstract: Globally, there is massive uptake and explosion of data, and the challenge is to address issues like scale, pace, velocity, variety, volume, and complexity of this big data. Considering the recent epidemic in China, modeling of COVID-19 epidemic for cumulative number of infected cases using data available in early phase was big challenge. Being COVID-19 pandemic during very short time span, it is very important to analyze the trend of these spread and infected cases. This chapter presents medical perspective of COVID-19 toward epidemiological triad and the study of state of the art. The main aim of this chapter is to present different predictive analytics techniques available for trend analysis, different models and algorithms, and their comparison. Finally, this chapter concludes with the prediction of COVID-19 using Prophet algorithm indicating more faster spread in short term. These predictions will be useful to government and healthcare communities to initiate appropriate measures to control this outbreak in time.

Parametric performance optimization of three sides roughened solar air heater

Abstract: Solar clean energy can be harnessed by several methods using technologies like solar heating, photovoltaic cell, solar architecture, photosynthesis. Solar energy is converted either by active techn...

Active Control of Uncertain Seat Suspension System Based on a State and Disturbance Observer

Abstract: In this paper, an active seat suspension system that is robust to uncertainties and unknown road profiles is designed using a state and disturbance observer. The controller which needs the measurement of just the seat frame position is analyzed with a 4-degree-of-freedom biodynamic model of the passenger for performance and stability. The proposed scheme is validated by simulation for a variety of road profiles under various uncertainties and compared with passive suspension, linear quadratic regulator-based active seat suspension, and conventional active cabin suspension using the evaluation norms set by ISO-2631. The proposed scheme is seen to be robust to uncertainties in seat frame suspension, passenger mass, and cabin suspension.

Data Analytics: COVID-19 Prediction Using Multimodal Data

Abstract: Globally, there is massive uptake and explosion of data, and the challenge is to address issues like scale, pace, velocity, variety, volume, and complexity of this big data. Considering the recent epidemic in China, modeling of COVID-19 epidemic for cumulative number of infected cases using data available in early phase was big challenge. Being COVID-19 pandemic during very short time span, it is very important to analyze the trend of these spread and infected cases. This chapter presents medical perspective of COVID-19 toward epidemiological triad and the study of state of the art. The main aim of this chapter is to present different predictive analytics techniques available for trend analysis, different models and algorithms, and their comparison. Finally, this chapter concludes with the prediction of COVID-19 using Prophet algorithm indicating more faster spread in short term. These predictions will be useful to government and healthcare communities to initiate appropriate measures to control this outbreak in time.

NegBERT: A Transfer Learning Approach for Negation Detection and Scope Resolution.

Abstract: Negation is an important characteristic of language, and a major component of information extraction from text. This subtask is of considerable importance to the biomedical domain. Over the years, multiple approaches have been explored to address this problem: Rule-based systems, Machine Learning classifiers, Conditional Random Field models, CNNs and more recently BiLSTMs. In this paper, we look at applying Transfer Learning to this problem. First, we extensively review previous literature addressing Negation Detection and Scope Resolution across the 3 datasets that have gained popularity over the years: the BioScope Corpus, the Sherlock dataset, and the SFU Review Corpus. We then explore the decision choices involved with using BERT, a popular transfer learning model, for this task, and report state-of-the-art results for scope resolution across all 3 datasets. Our model, referred to as NegBERT, achieves a token level F1 score on scope resolution of 92.36 on the Sherlock dataset, 95.68 on the BioScope Abstracts subcorpus, 91.24 on the BioScope Full Papers subcorpus, 90.95 on the SFU Review Corpus, outperforming the previous state-of-the-art systems by a significant margin. We also analyze the model’s generalizability to datasets on which it is not trained.

Proposing building information modeling-based theoretical framework for construction and demolition waste management: strategies and tools

Abstract: Waste generated throughout the construction and demolition (C&D) process has become a great challenge and hindrance toward the sustainable development as it creates environmental degradation. Prope...

Virus Graph and COVID-19 Pandemic: A Graph Theory Approach

Abstract: In the field of science and technology, the graph theory has offered several approaches to articulate any situation or concept. The use of graph theory enables the users to understand and visualize the situations like COVID-19. Looking at this pandemic disease, its impact and the preventing measures, the graph theory would be the most appropriate way to exercise the graph models with theoretical as well as practical aspects to control this epidemic. In the context of COVID-19, this chapter defines the variable set, variable graphs, and their types considering the variations in the vertex sets and edge sets. The virus graph and their type are discussed in this chapter that states that the Virus graph type I and III are not so perilous for all living beings, but virus graph type III and IV are extremely hazardous for the harmony of the world. Initially, the COVID-19 was in Virus graph-I type, but presently it is in Virus graph-II type. Given different aspects for expansion of pandemic, this chapter presents growth types of virus graphs and their variation as 1-1, 1-P, and 1-all growth types. This chapter provide the number of infected people after ‘n’ number of days concerning different values of P and growth rates with I0 = 100. At the end of this chapter, the country-wise starting dates of stages of the virus graph-I and II are specified. The concept of cut sets is applicable for the prevention of COVID-19 and the whole world is using the same analogy.

Feature selection using principal component analysis and genetic algorithm

Abstract: Feature engineering is the way toward utilizing domain knowledge of the records to build features that in turn assist Machine Learning (ML) algorithms to provide efficient results. It is crucial to...

Detection of COVID-19 from CT Images Using resnet50

Abstract: Coronavirus (COVID-19) outbreak which is a global health pandemic occurred due to (SARS) Severe Acute Respiratory Syndrome and SARs-CoV-2 has an immeasurable im

Organic Surfactant Assisted Polypyrrole Materials as Effective Chemiresistive Gas Sensors for VOCs and Toxic Gas Detection

Abstract: Surfactant assisted PPy materials are explored as efficient gas sensors for VOCs and toxic gases in the present work. Polypyrrole nanostructures were prepared by chemical oxidative polymerization using various sulfonic acids as surfactants between 0–5 °C. The effect of sulfonyl group on PPy was confirmed by FT-IR spectrum analysis. Doping of sulfonic acid has a strong effect on morphology of polymer blends which displayed the amorphous, globular, cauliflower, flakes and compact sheet like structures observed in the surface morphology. The conductivity of doped PPy observed to be dependent on the nature of dopant and its doping level. The electrical conductivity was observed to enhance by an order in magnitude on the addition of organic surfactants from 0.1 to 2.0 M. The pristine PPy shows high % response of 93% and 70% only for NH3 and Cl2 respectively. But different surfactant doped PPy were found to be highly sensitive for most of the VOCs and toxic gases which emerged as efficient sensor materials for fast response and recovery at the lowest level of 5 ppm.

Impact of 2018 Kerala Floods on Soil Erosion: Need for Post-Disaster Soil Management

Abstract: Extreme precipitation events lead to flash floods, which can trigger soil erosion and landslides. While damages to infrastructure and livelihoods are rapidly assessed on economic terms, damages to natural resources are not estimated due to limited observation record. This study conducted an analysis using remote sensing data to estimate changes in soil erosion rates before, during and after the Kerala 2018 floods, based on the Universal Soil Loss Equation (USLE). The USLE was driven by multiple data including: in situ rainfall data from Indian Meteorological Department (to estimate rainfall erosive factor), soil maps prepared by Food and Agriculture Organization (to estimate the soil erodibility factor from the properties of soil that consists of the percentage of clay, loam and silt), digital elevation model (to estimate topographic slope and length) from Shuttle Radar Topography Mission and multispectral imagery (to estimate cover management factor and conservation practice factor) from Landsat-8 satellite. Data from these sources were analysed using a Geographic Information System (GIS) platform. Results indicate a state-wise average increase of 80% (31–56 metric tons ha−1 year−1) in soil erosion rate during the floods. Of the districts, Idukki showed the highest increase, of 220% and more susceptibility to soil erosion, which is in comparison with government survey records. Results show that the floods and associated erosion were not only due to the rainfall event but also due to the rapid change in land use and land cover, from natural to human settlements. Therefore, government agencies need to protect land cover and reduce unsustainable development in ecologically sensitive environments, which if managed properly can act as a buffer for soil erosion extremes in Kerala.

Control of Antilock Braking Systems Using Disturbance Observer With a Novel Nonlinear Sliding Surface

Abstract: In this article, the problem of control of antilock braking systems under significant uncertainties and unknown relationship between tire-road friction coefficient and wheel slip ratio for a two-axle vehicle model is considered. Two methods based on a disturbance observer and sliding mode control are proposed. In the first method, a conventional sliding surface is used while in the second method a new nonlinear sliding surface is proposed as an improvement over the first method. The performance under the proposed methods is assessed analytically and by MATLAB simulation under different road conditions. The proposed methods are validated further on CarSim platform.

Image Retrieval Using Low Level and Local Features Contents: A Comprehensive Review

Abstract: Billions of multimedia data files are getting created and shared on the web, mainly social media websites. The explosive increase in multimedia data, especially images and videos, has created an issue of searching and retrieving the relevant data from the archive collection. In the last few decades, the complexity of the image data has increased exponentially. Text-based image retrieval techniques do not meet the needs of the users due to the difference between image contents and text annotations associated with an image. Various methods have been proposed in recent years to tackle the problem of the semantic gap and retrieve images similar to the query specified by the user. Image retrieval based on image contents has attracted many researchers as it uses the visual content of the image such as color, texture, and shape feature. The low-level image features represent the image contents as feature vectors. The query image feature vector is compared with the dataset images feature vectors to retrieve similar images. The main aim of this article is to appraise the various image retrieval methods based on feature extraction, description, and matching content that has been presented in the last 10–15 years based on low-level feature contents and local features and proposes a promising future research direction for researchers.

An experimental-based python programming for structural health monitoring of non-engineered RC frame

Abstract: Most of the damages were experienced on the buildings which were conventionally built without any consideration of IS codal provisions conveniently called non-engineered structures. Non-engineered structures are frequently affected by vibrations due to various natural and artificial sources. Thus, it needs special attention. It is, therefore, necessary to check the performance of non-engineered structures through various health monitoring techniques. A piezoelectric-ceramic (PZT) sensor-based technique called electromechanical impedance (EMI), in which the sensors efficiently operate at a high-frequency range and can typically detect damage at the initial level which is implemented for the purpose. In this research work, experimental tests are performed on the non-engineered reinforced concrete frame using EMI technique by utilizing a PZT sensor which is bonded to the structure using the high-strength epoxy adhesive. The experiment is carried out to identify and locate the damages using frequency variations, and the severity was checked using extracted equivalent parameter; damage index. Second, a Python programming is developed by the authors to identify and quantify the damage index and root mean square deviation index in the frame. The frequency responses obtained from the experimental tests are used in the programming. The performance of the program is compared with the experimentally calculated parameters to check the efficiency of the programming. According to the results of the comparison, it is observed that python programming can be effectively used for damage detection.

Novel Face Liveness Detection Using Fusion of Features and Machine Learning Classifiers

Abstract: As the technology is growing which is birth to different types of frauds into areas like face detection, finger print detection etc. Moreover, it becomes very difficult for service providers to maintain security of data. In addition, these systems needs to be protected from spoofing by attackers. Fraudsters can use dummy Eyes, photographs for identification of faces for authentication purposes. These facial recognitions can also be done through face detection from video streams by replaying videos & capturing specific moments of person. Also these kind of attacks can be done successfully as system can't detect the real life faces & faces extracted from videos & photos. In addition, these videos & photos will be easily available on internet & other stored media, so anyone can challenge the system by using it. Now, many algorithms are implemented to detect face liveness for authentication. The paper represents novel face liveness detection using fusion of luminance-based features with help of assorted machine learning classifiers.