Proceedings of Fourth International Conference on Communication, Computing and Electronics Systems

Conical tank process (CTP) is used in most of the industries, namely food processing, chemical, beer production, pharmaceutical, and waste treatment plants. This unit is widely used because of its accurate and thorough mixing capability. Efficient operation of these industries would require control of enormous process variables within their specified bounds. Today, control of large continuous interconnected plant is a challenging task and needs complex control hardware. Although the conventional PID control algorithms are widely used, modern control method such as adaptive control can provide noteworthy improvement. CTP has nonlinear characteristics due to its variable area and control of level has become an exigent task. Most of the industrial process loops are interacting in nature and require enormous control effort to achieve desired level. In order to accommodate changes in process parameters, gain-scheduled PI controller (GSPI) is designed. A major implementation issue of GSPI is design of the switching function to have a smooth changeover in process variable according to the operating point variations. A fuzzy logic system is used to overcome this difficulty. In fuzzy gain-scheduled PI (FGSPI) controller, based on expertise of human operator fuzzy rules are framed, and with the associated reasoning mechanism, the controller parameters are determined. In this paper, the elucidation of proposed FGSPI controller is simulated on a interacting conical tank process (ICTP). A simulated result shows the effectiveness of FGSPI controller in terms of setpoint tracking, disturbance rejection and process parameter variations.

Direction of arrival (DOA) estimation plays a critical function for beam forming in the massive multiple-input multiple-output (mMIMO) 5G wireless systems to increase the coverage, capacity and throughput. Accurate direction of arrival estimation of the received signals from large number of user equipment is a major issue in millimeter wave (mmWave) wireless systems due to multipath signal propagation with reflections. In this paper, we review the conventional and the state-of-the-art deep learning-based direction of arrival estimation techniques and compare with conventional techniques in the context of high resolution and accuracy.

In digital signal processing applications, the critical route includes the parallel summation of multiple operands. High compression ratio counters and compressors are required to speed up the summing operation. In the proposed work, (15,4) counters and (7,3) counters are designed, while the counters are sorted using the proposed sorting network. The counter's inputs are split asymmetrically into two groups and fed into sorting networks to produce reordered sequences that can only be represented by one-hot code sequences. The counters are constructed in Xilinx using Verilog and obtained the results of time and delay. This is then applied to the sorting network and checked the sorting network for counters (7,3) and (15,4). In addition to this, sorting network was made effective in sorting the larger numbers either in the ascending or in the descending order.

In the case of low resource language, there is still the requirement for developing more efficient Automatic Speech Recognition (ASR) systems. In the proposed work, the ASR system is developed for the Gujarati language publicly available dataset. The approach in this paper applies the combination of Mel-frequency Cepstral Coefficients (MFCC) with Constant Q Cepstral Coefficients (CQCC)-based integrated front-end feature extraction techniques. To implement the backend part of the system, hybrid acoustic model is applied. Two-dimensional Convolutional Neural Network (Conv2D) with Bi-directional Gated Recurrent Units-based (BiGRU) backend model is used as the model. To build the ASR system, Connectionist Temporal Classification (CTC) loss function, CTC and prefix-based greedy decoder are also used with the acoustic model. The proposed work shows that the joint MFCC and CQCC feature extraction techniques show the 10–19% improvement in Word Error Rate (WER) as compared to isolated delta-delta features with the available integrated model.

Right from the discovery of electricity, human fraternity is searching for efficient techniques to store that energy in order to meet the future energy demands. Recently, the energy storage engineering is evolving by adapting innovative technologies and trying to meet out the varying energy requirements. Hence, identifying suitable, sustainable, environmental-friendly and competent energy storage devices and materials has become the most important need of the hour. Nanotechnology and nanomaterials have extremely decisive responsibility in the present energy storage sector. Nanomaterials are considered to be the “wonder” materials with their remarkable characteristics such as fast ion diffusion, high particle volume, better electronic conductivity make them as competent nominee for energy storage. The surprising properties of carbon-based nanomaterials and their tunable surface chemistry permit them as the support for high-power energy storage devices. This article traces the role of few carbon-based nanomaterials, for instance, graphene and fullerenes in practically influencing and improving the ability and dependability of devices used for energy storage like batteries, supercapacitors and fuel cells.

Due to innovative growth in cyberspace and portable penetration, kids in India are at risk of cyber harassment. A review of 174 central graders in Delhi exposed that a total of 8% pandered in cyber harassment and 17% described being offended by such acts. However, the occurrence of in-person harassment, hostility, and discrimination is also happening. Amid all the digital platforms, Instagram positions developed with extreme cyber harassment. Over the preceding span, it brought substantial developments in the grounds of machine learning (ML) which have been efficaciously functional in fields associated with cyber harassment findings, such as buzz recognition, sentimentality study, and forged broadcast discovery. In machine learning, methods are provided for detective work and the classification of online harassment. Researchers have conducted experiments using comparable datasets. To build an ensemble learning model for detecting and classification different categories of online harassment from social media platforms. In our proposed work, a robust method of detecting online harassment (cyberbullying) on the Instagram dataset is used. The attributes of abusive words are initially analyzed from feature selection and pre-trained word embedding language models like BERT and ELMO. The harassment words are detected using unsupervised machine learning techniques such as association rule classifier, latent semantic analysis (LSA), and clustering technique. Then, a novel ensemble typical model is planned for categorizing the different types of online harassment using extreme gradient boosting (XGBoost) learning method. Hence, our robust method of ensemble model for detecting and classification of online harassment provides much better results with high accuracy and lesser loss function.

This paper helps in automating process of car parking in shopping malls. It helps in making parking more efficient by burning of less fuel. This system is useful for places with large number of people considering less people-to-people contact considering Covid Pandemic and making a safe system for minimal infection transmission from people to people. This paper aims at developing a IoT-based E-parking system. This project uses Micro-controller (ATtiny85) for controlling of sensors. Set of multiple ultrasonic sensors are put on ceilings per floor with multiple slots for detection of vehicles in parked spaces with threshold set for cars. Multiple Wi-Fi modules are used for wirelessly uploading the values of vehicles parked in different floors to cloud from where the Wi-Fi module at entrance extracts data and displays on central display at entrance for assigning empty parking slots to new vehicles on arrival. Entrance display displays number of empty slots on every floor to new customer entering mall parking system. This project achieved objective of making a system which can be used in times of Covid-19 for better safety of people. This paper has been able to achieve its main objectives of making a safe, affordable, scalable parking system which can be used in shopping malls and multiplexes. It can be scaled to large usable parking systems using better sensors and better computing devices. It can provide means of work or business to youth of city for building and selling smart vehicle parking systems and deploy them to multiple malls and multiplexes using help from staff and sell at affordable rates. It can also help make more customizable and modular smart parking systems tailored to use of system in any buildings. Arduino IDE has been used for uploading code to cloud modules in project.

The Advanced Encryption Standard (AES) is the most widely used symmetric and secure encryption algorithm for commercial and research purposes. The core of AES is its substitution box (S-box), which primarily provides nonlinearity, confusion, and diffusion. The improvements to nonlinearity in S-box is one of the methods used by researchers to enhance security in AES. In this paper, to increase the nonlinearity of the S-box, AES using a key-dependent dynamic S-box is implemented and compared with the AES using a static S-box. The algorithm proposed in this paper draws observation from results of security parameters like hamming distance, strict avalanche criteria (SAC), and balanced output while comparing key-dependent S-box and static AES S-box. Among 40 experimental results, the key-dependent dynamic S-box performs 22.5% times better for hamming distance observation drawn is that the hamming distance is more in the dynamic AES than static AES. The average bit difference of 10 rounds of dynamic AES is around 65%, which is a good avalanche criterion.

Effective configurable adder is a traditional method used for obtaining high accuracy. In day-to-day life, everyone’s goal is to reduce delay and increase speed. The proposed study is mainly focused on these parameters for obtaining high accuracy, less power, and speed operation. The proposed adder includes an effective configurable adder with high-speed error detectable model with easy testability. By incorporating the process of traditional carry look-ahead adder, the proposed adder utilizes the propagation with carry masking method. Then, the accuracy has been tested to remain configurable at run-time. The results from the proposed study indicate that the implementation of proposed design on a Spartan3E FPGA utilizes 62% fewer slice registers and 47% fewer slice LUTs when compared with the standard configurable adder design.

Since the start of the industrial revolution from the fifteenth century man-made machines and the usage of fossil fuel paves the way for today’s air pollution. Air pollution leads to the death of nearly seven million people all-round the world. Air pollution means the release of carbon dioxide, sulphur dioxide, nitrogen dioxide, methane and particulate matter which includes PM2.5 and PM10 in the atmosphere which are the major metrics for measuring the air quality index (AQI). In addition to these pollutants, due to the increased production and improper disposal of electronic goods, several other poisonous air pollutants such as chromium dioxide impact the quality of breathable air severely. This article addresses the issues caused by improper handling of the e-waste and also proposes a new way to monitor and instantly report the air quality around the e-waste dump yards with the help of IoT in the Vellore region.

Unnatural fatalities are rising in our contemporary period, but road accidents account for a significant portion of these deaths. This work mostly deals with road accident detection and methods to reduce the time period between the victim and the emergency service area. According to the statistics, some accidents have occurred due to some momentary health issues like low/high BP, heart rate abnormality, lack of intaking oxygen-increased breathing rate, and also due to drunk and driving, these kinds of stuff are incredibly hazardous, resulting in automobile collisions and traffic injuries. So, continuous passenger and driver monitoring was also introduced for the utmost safety. Besides of preventative actions implemented, like seizure of vehicle permits, fines, penalties, and seizure of licenses. Despite the numerous prolepses taken, the number of accidents caused by drunk driving and sudden heart attacks is on the rise. This module is suggested to prevent individuals from dying needlessly as a result of intoxicated driving incidents, momentary health issues with the help of continuous monitoring, and to detect the occurrence of accident and also sends location coordinates to nearby hospitals/toll plazas/police stations and passenger’s trustworthy relatives for their navigation purpose. This module consists of Raspberry Pi 0, Arduino UNO, alcohol detection sensor (MQ-2), MEMS Accelerometer Sensor, ECG, GPS module, pulse oximeter, digital vibration sensor, LCD displays, ADC, PHP server, and relay to control the vehicle.

In today's world, acts of abducting and brutality against people are increasing at an alarming rate. According to World Health Organization (WHO) poll, 35% of individuals are abducted around the world. Even though smartphone technology has advanced rapidly, holding phone to make a call or send message remains inconvenient. As a result, Protecta watch was created, which helps to prevent danger. It is an IoT device that transforms smartwatch with a safety device. It generates a live location sharing, an alert message, and an emergency call dial that the victim starts with a click of a button. It uploads data of mobile global positioning system (GPS) to the cloud every 5 min using Blynk server. The circuitry is designed on a printed circuit board (PCB) for a form factor. This device is both affordable and revolutionary in terms of protection strategy due to the convergence of all three functional areas.

The introduction of Internet of Things (IoT) has benefitted smart cities, healthcare, and transportation applications for enabling smart connections and analyzing the performance. As the number of IoT devices and potential connectivity increases, the distance between smart hubs, smartphone software complexity, and people behaviors changes rapidly resulting in a power utilization challenge. Energy efficiency and power dissipation are two of the most significant challenges in green IoT-enabled technologies. Based on a systematic literature review, this study presents the power management possibilities in the Internet of Things (IoT) domain. The main purpose of this study is to develop energy-efficient approaches for WSNs that enable IoT. The “Green IoT” concept aims to reduce the energy consumption of IoT devices while maintaining sustainability conditions. In this study, which is driven by establishing a viable IoT environment for the future phase and moving the globe toward the formation of Green IoT, a review of Green IoT (GIoT) is first provided, and then the challenges and potential possibilities for GIoT are studied.

Plants can be classified based on various classification methods such as cell, genetic and serum etc. It's difficult for an individual to explore the various classification methods and it's practically not feasible as it demands good knowledge in plant taxonomy and long-term time investment. Due to the shortage of experienced and qualified taxonomists in identification and classification of medicinal plants, with the help of different image processing algorithms and computer vision, the above difference can be bridged. The main objective is to develop a Deep Learning and Machine Learning based model to identify and classify plants based on various features, which is done with the help of Gabor filter and Gray Level Co-occurrence Matrices (GLCM) and using classifiers such as Random Forest (RF), and Light Gradient Boosting Machine (LGBM), and made a comparative analysis which resulted an accuracy of 95.5% with LGBM and GLCM filter and used to develop a standalone device that clicks a picture and identifies the medicinal plant.

Filters are an integral part of Digital Signal Processors (DSPs), which are necessary for signal processing in digital devices. DSPs are required to process sound signals to provide meaningful data and they are also required to process images captured by digital cameras. Multipliers and adders together form most parts of a filter. Existing filter architectures utilize various multipliers like Array multiplier, Wallace multiplier, Vedic multiplier, etc. These multipliers have a high delay resulting in performance degradation of the filters. Thus, an improvement in the architecture of the multiplier results in the betterment of the overall performance of the filters. The multipliers in turn can be improved by making changes in the algorithm used for the reduction of partial products and also by making changes to the architecture of the adder. For this purpose, a modified Dadda multiplier in which the final addition is done using the Kogge-Stone adder is being proposed in this work. The designed multiplier is an 8-bit multiplier that takes in 9-bit signed magnitude values as input and gives out 23-bit outputs which are then added together for the output to be produced. The FIR and IIR systems have been designed using VHDL and implemented in Vivado 2017.4 to obtain results.

In most of the prominent areas of digital signal processing (DSP) applications, the operation of linear convolution finds a pivotal role. With the increasing demand for applying convolution to varied applications of DSP, there has also come a need to develop techniques to compute complex high-speed computations. Hence, this paper focuses on providing a linear convolution architectural model for computing the operation fast and accurately with minimal power dissipation and area. Here, the 16-bit linear convolution architecture has been designed by proposing a set of 4-bit reversible Vedic multipliers (RVMs) architecture, 8-bit and 9-bit reversible Kogge–Stone adders (KSAs). The 4-bit RVM architecture is built based on the Urdhva Triyagbhyam (UT) algorithm and the 8-bit and 9-bit KSAs employ the proposed reversible black cell (PRBC) and proposed reversible grey cell (PRGC). The performance of the proposed reversible KSA is compared with existing reversible KSA and other parallel-prefix adders (PPAs), designed using PRBC and PRGC. The PRGC and PRBC have resulted in a nearly 50% reduction in the number of clock cycles (CCs), constant inputs (CIs), and garbage outputs (GOs) compared to the existing grey and black cells. The 16-bit linear convolution system has been designed using VHDL and implemented in Vivado 2017.4 to obtain the results of power, area, and delay.

Globally, the growing number of elderly people, chronic disorders and the spread of COVID-19 have all contributed to a significant growth of Home Health Care (HHC) services. One of HHC’s main goals is to provide a coordinated set of medical services to individuals in the comfort of their own homes. On the basis of the current demand for HHC services, this paper attempts to develop a novel and effective mathematical model and a suitable decision-making technique for reducing costs associated with HHC service delivery systems. The proposed system of decision making identifies the real needs of HHCs which incorporate dynamic, synchronized services and coordinates routes by a group of caregivers among a mixed fleet of services. Initially, this study models the optimization problem using Mixed Integer Linear Programming (MILP). The Revised Version of the Discrete Firefly Algorithm is designed to address the HHC planning decision-making problem due to its unique properties and its computational complexity. To evaluate the scalability of this proposed approach, random test instances are generated. The results of the experiments revealed that the algorithm performed well even with the different scenarios such as dynamic and synchronized visits. Furthermore, the improved version of nature-inspired solution methodology has proven to be effective and efficient. As a result, the proposed algorithm has significantly reduced costs and time efficiency.

In the present scenario, Handwriting Recognition (HWR) plays a very important role as due to current pandemic situations most of the exams are conducted in online mode. In HWR the device translates the user’s handwritten characters or words into readable by a computer system. The problem with HWR is there are different types of handwriting and different styles of writing each character which makes it difficult to analyze each letter uniquely and correctly also the major problem with HWR is when people write there is no specific font and font sizes are taken care which also plays important role in recognizing the letters. The paper presents an accuracy comparison system of spelling correction in Handwritten OCR data using four neural models BERT, SC-LSTM, CHAR-CNN-LSTM, CHAR-LSTM-LSTM. In task, sequence matcher algorithm is used for computation of similarity score between input text data and outputs of the different neural models at each iteration. It is observed that the BERT model gives the highest accuracy of 71.4% followed by SC-LSTM with 69.12%, CHAR-CNN-LSTM with 67.80% and CHAR-LSTM-LSTM with 69.34%.

The term “augmented reality” refers to a technology that combines digital and actual experiences. It is an immersive experience of a physical environment in which actual objects are enhanced with digital visual features, sound, or other sensory stimuli. The rapid advancement of augmented reality has piqued people’s interest in recent years. It is a rapidly developing area among businesses that deal with mobile computing and commercial apps. Using AR, digital information can be placed in reality to improve a human’s perspective of reality. This paper begins by defining augmented reality, its history, and its challenges. The paper then discusses some essential technology, development tools, and augmented reality applications in several industries. The main focus point of the paper is centered around the discussion on EDGE Technology as a solution to the limitations of AR. We have drawn a comparison between some frameworks that have been developed over time, merging AR with EDGE. Finally, it anticipates future advancements in augmented reality technologies, such as the Mobile AR.

Face recognition-based phone unlocking was first introduced by Apple on its iPhones in 2017 and has since become a disruptive breakthrough in the smartphone business. Home security, monitoring, and automation technologies, meanwhile, have recently become an important part of many people’s daily lives. This Face Recognition and Voice Recognition Smart Door Unlock System aims to increase security. In this system, a camera sensor is utilised to capture the face, and an image matching algorithm is used to recognise authorised faces, as well as a voice sensor to record sound. Only the person with the matching face or voice may unlock the door. The security system is also built by including maintenance into the design.

Artificial intelligence’s value has increased in recent years. Artificial intelligence (AI) backed by big data analytics has expanded over the past few years. According to reports and reviews, artificial intelligence structured on large volumes of data analytics and information and communications technology has the potential to greatly improve supply chain performance; however, research into the reasons why companies engage in manufacturing activities and the novel artificial intelligent systems is limited. It is in this regard that this study has been carried out. To this end, several theoretical approaches have been proposed as explanations for how manufacturing businesses generate valuable resources and worker skills to impose innovation and enhance circular economy proficiency. The goal of this study is to gain approval for an intellectual concept that explains how institutional pressures on resources affect the implementation of big data in artificial intelligence, as well as its influence on sustainable manufacturing and the model of production and consumption proficiency when regulating the effects of industrial flexibility and industry effectiveness. We believe that if companies want to see a meaningful return on their AI efforts, they must fill this gap and promote AI capability. It is on this central aim that this study will expose and encourage research into this area; moreover, it hopes to create awareness among new industrial facilities of the essence of implementing AI features to boost any form of manufacturing and fabrication process.

For performing efficient modular arithmetic operations, several cryptographic and pseudorandom bit generator (PRBG) algorithms utilize a 3-operand binary adder as the primary functional unit. The Carry Save Adder is the most common adder used for performing the three-operand extension (CS3A). On the other hand, the ripple-carry step of CS3A results in a significant delay while transmitting the output signals. Due to the lengthy delay, it influences the performance of MDCLG architecture. For performing three-operand addition, two-operand adders, such as Kooge Stone (KSA), can be used. This will decrease the critical route latency, delay, and area compared to other parallel prefix adders. The proposed high-speed and space-efficient adder architecture for performing three-operand binary operations includes carry-prefix computation logic after performing the pre-compute bitwise addition. The proposed adder design reduces the adder latency while consuming less area and power. A Kogge–Stone parallel prefix adder has been used to develop a novel architecture for the proposed 8-bit, 16-bit, and 32-bit three-operand adders. The proposed architecture is implemented by using Verilog coding, and further, the power and delay extraction has been performed by using a Xilinx tool. The proposed architecture has been developed by using the MDCLCG method with the three-operand adder, and further, the proposed architecture is proven with respect to delay as well as area and power.

Mulching is a long-standing agricultural practice that entails spreading a layer of organic matter around plants to shield their roots from heat, cold, or drought, as well as to keep the crop clean. Mulching provides its own set of benefits like retention of soil water content, protecting the soil from erosion, harsh winds, hot sunshine, and overall damage due to other environmental and external factors. In recent decade, the demand for mulch laying machine has been drastically increased with the abovementioned benefits. However, it is a very labour-intensive process if done manually, thus the need for an automatic one. In this project, we have prototyped an automatic mulching machine. Initially, a software model was designed using Autodesk Fusion 360. The hardware modelling was done with the help of an Arduino, a Bluetooth module, slider-crank mechanism, and the interface of hardware with software was done using a mobile app.

The mmWave technology is employed because it has a big bandwidth and is designed with hybrid beamforming using a huge antenna array. The amount of BF gain required for effective transmission is determined on how far away the user is from the base station. The user’s distance affects the beamforming (BF) gain. The amount of antenna elements affects BF gain. So, hybrid beamforming (HBF) design user needs are met. Because certain groups of users are supplied by particular groups of RF chains, RF chains are grouped. Using link adaptation techniques, the data rate may also be boosted. Simultaneously, the required bit error rate (BER) is obtained. As the channel has nonlinear features, it is dependent on specified signal-to-noise ratio (SNR) threshold values, which affects the system’s performance. As a result, we use link adaptation approaches in two stages. In diverse places, data streams are represented utilizing wireless sensor nodes. During the beginning phase of the link adaptation process, the digital precoder and combiner are switched between modes according to the characteristics of the channel. A machine learning (ML)-aided link adaptation is applied in second step. The base station (BS) receiver anticipates whether it should request spatial multiplexing or diversity-aided transmission for each additional channel realization. Both the precoder and the combiner are unnecessary for a single dominating route. Most notably, learning-aided adaptation produces more data than conventional link adaptation.

Alzheimer’s disease is an elderly chronic disease, which affects the people with age more than 60. In India more than 5 million people are affected by the Alzheimer’s disease (AD), it may be increased to 7.6 million by 2030. 6.2 million US people are living with AD and may be increased by 8.6 million by 2030. As per a report of Alzheimer’s Association in worldwide every 10 of 100,000 people are developing AD each year. As per the report of world health organization nearly 55 million people are living with AD in worldwide and for every year 10 million new cases are developed. By 2022 the cases may be increased to triple the count of now. This is the 7th leading death causes disease in world. If a person is affected by AD then they cannot able to do any activity by their own because of the memory loss. The caretakers will guide them to their day-to-day activity. There are no medicines for the AD so we are in the situation of protecting the people from this disease. In this paper the early prediction has been focused. Early prediction of AD and the diabetics has been done with the retinal fundus image. The nerves in the brain resembles in the retina image. The image preprocessing is done and the image is segmented. The segmented image is given to the deep learning model for the prediction of the disease. The accuracy of this model is 90%.

Plant diseases could lead to huge production loss for the cultivators. These diseases are typically in the form of visible symptoms like color changes on the surface of the leaves, different colored spots, or streaks. This region of interest is extracted using image processing, and the area of the disease-affected part of the leaf is calculated. This system is proposed to support agriculturists to identify plant diseases efficiently and constantly monitor the health conditions of the plants. A convolutional neural network is used to identify common diseases of a few types of fruit leaves. The overall accuracy of this system is found to be 90% with a loss of 2.8%. Determining the disease and the leaf’s disease-affected area will help in maintaining a better quality of the crop by taking the required actions.

The recent trends in the gaming world have been more inclined to shooter games which covers a wide range of audiences including streamers and many more people. People expect the games to be closer to reality for a lively experience. Behaviors of non-playable characters (NPC) in various games like Grand Theft Auto are studied and various methods of defining behaviors to non-playable characters like behavior trees and Q-Learning behavior tree are compared for performance and activities. The comparisons of the resultant agents are made using certain performance criteria like the closeness of AI behavior to the humans, latency to respond to events etc.,

Significant gains in semiconductor technology devices have enabled high-data-rate communications at 60 GHz which stimulates short-range multigigabits-per-second transmission for multimedia applications. This work focusses on the design of the first block of the receiver and the low-noise amplifier with integrated antenna which is considered to be the most challenging task. Over the desired frequency of 60 GHz, microstrip antenna and low-noise amplifier have been designed and integrated with co-design approach. Using inductive source degeneration technique, two-stage common source low-noise amplifier in a 65-nm CMOS technology has been designed and found to produce gain of 12.557 dB and noise figure of 3.626 dB. Antenna efficiency is the amount of RF power delivered to the antenna (from radio) which is actually transmitted into the air.

High Data Dimensionality Reduction (HDDR) removes the irrelevant features in a complex dataset and incorporates various techniques that could be used to foretell the research outcomes in a predictive model. The major objective of the paper is to analyse and survey the various models on the basis of HDDR and its feature pre-processing methods applied in financial dataset predictions. Numerous techniques of data mining and its strategies were discussed and assessed to ascertain the importance of augmenting the performance of the financial dataset with classifiers. The pre-processing techniques applied in various research works and their outcomes are highlighted. The HDDR methods used in financial prediction of Small Medium Enterprises (SMEs) are studied for existing frameworks and models by different authors. The paper encapsulates the gist of the models, frameworks and algorithms involved in effective elimination of irrelevant features and extraction of best features for best prediction of financial datasets.

Network analysis plays a significant role in business which is achieved through community detection. The relationship between the nodes is mined by community detection, which facilitates the analysis of complicated networks. In the current era, social media is the commonest mode of communication that leads to complex online social networks, from which useful information can be retrieved. Not only textual data is shared in social media, in addition multimedia data plays a significant role in content sharing in social network. This research work proposes a community detection method based on deep learning algorithm to detect communities for both textual content and multimedia content. The proposed method triphase–deep learning community detection (TriDL-CD) method illustrates the relationship of data in a graphical way in the first step. Second step converts the graph into user relationship table using similarity weightage analysis from which the communities are formed using convolution neural network. The proposed method proved to be efficient in detecting high quality communities for multimedia content compared with Louvain community detection algorithm, Leiden community detection, and surprise community detection algorithms. Experimental results show the efficacy of deep learning in the concept of community detection.

Data security is becoming a major concern in recent years due to the proliferation of interconnected devices. With the number of devices connected to the Internet of things (IoT) growing at a very rapid pace, design of speed and power efficient crypto systems is very essential. This paper explores the design of a high-speed and low-power advanced encryption standard (AES) architecture to be used as a hardware accelerator in various cryptographic systems. The use of Rijndael S-Box for byte substitution, merging of ShiftRows operation with the State Register updation, implementation of MixColumns transformation through substructure sharing, and clock gating methods are investigated in this work for obtaining a high-speed device with low-power consumption. The proposed design is designed using Verilog and co-simulated in MATLAB and implemented as an ASIC using 90 nm GPDK technology. It is seen that the proposed design has 83% increase in speed, 39% decrease in area, and 82% reduction in power dissipation as compared to a standard AES design implementation.

Obstructive sleep apnea (OSA), a sleep condition, is characterized by recurrent bouts of irregular breathing. This study uses deep learning (DL) and variational mode decomposition (VMD) techniques to develop a classification system for sleep apnea. VMD is an adaptive signal decomposition technique to simplify complicated signals into a finite number of decomposed intrinsic mode functions (IMF). The baseline systems consist of support vector machine (SVM)-classifiers constructed using statistical features. Each of the ECG recordings is segmented; into one-minute segments. VMD is then performed on each of the one-minute long ECG segments. Statistical features derived from the derived IMFs, time, and frequency domain (TD and FD) segments are used to train SVM classifiers. We then developed convolutional neural networks (CNNs) and evaluated the performance. The CNN models trained with first, second, and third IMFs were the best performing systems. Subsequently, we added the first, second, and third IMFs to reconstruct a denoised ECG signal. CNN model trained with this regenerated ECG signal showed an accuracy, sensitivity, and specificity of 88.187%, 93.128%, and 80.339%, respectively. Subsequently, we extracted bottleneck features (BNF) from the bottleneck layer of the CNN. We trained a smaller dense neural network using the BNFs. When compared to the CNN model, the dense neural network trained with BNF extracted from the regenerated ECG signal (resulting from summing first, second, and third IMFs) gave the best performance with 4.24% 6.84%, and 2.55% improvements in accuracy, specificity, and sensitivity, respectively; with an area under the ROC curve 0.914.

In this work, a circular shaped ethanol-filled hollow core photonic crystal fiber (PCF) is proposed. The optical properties, like dispersion, effective area, confinement loss, and nonlinear coefficient of the proposed hollow core PCF, have been studied for the wavelength range from 800 to 1600 nm. The main focus of this research work is to achieve nearly zero dispersion wavelength (ZDW) by using finite element method. When air is poured into the entire hole ring, ethanol is poured into the middle hole ring, and ethanol is poured into the entire hole ring, then the ZDW of 880 nm, 1220 nm, and 1250 nm is achieved, respectively. This kind of PCF is useful in sensing applications, nonlinear applications, laser technologies, and telecommunication.

Indian electricity industry is under transition from vertical to restructured structure. In a 24 h electricity market, the operator performs auction for each hour in a day. Each time slot corresponds to different market clearing price. Energy consumption cost of intensive users can be reduced through demand response programs that involve planning of various processes stages in industries. This requires prediction of market price. This research work presents demand management planning model based on predicted price for a 24 h day ahead Indian electricity market. The proposed system is employed on the cooling production and distribution segment in a meat industry to control the chillers and dryers. The system employs autoregressive integrated moving average model to predict the Indian market price using real-time data. The ARIMA (2, 1, 5) appears to be an adequate model. The proposed model provides an evidence of significant saving in consumption cost in a day.

In this survey, the comparison results of current mode logic styles such as MOS Current Mode Logic (MCML), Dynamic Current Mode Logic (DyCML), and Positive Feedback Source Coupled Logic (PFSCL) gate structures are analyzed. In this, MCML and PFSCL are static logic circuits. The dynamic logic uses a clock signal as one of the inputs. The simulation results are performed at a voltage of 1 V and a temperature of 27 °C. The values of power, propagation delay, and power delay product are obtained and analyzed using the Cadence Virtuoso tool. The power and the delay values are verified with Monte Carlo simulations using a histogram plot of 200 samples. The process variations for different corners are simulated and the parametric analysis with different temperatures are compared for the different topologies of current mode logic gates. From the comparison, it is clear that Dynamic CML provides high performance and operates in a low-power environment.

Intelligent transportation has provided tremendous convenience to drivers as they have gained interest in Vehicle Ad Hoc Networks (VANETs). The messages can be broadcasted by any user on the open communication network. In some cases, the user can misbehave by sending wrong messages, malicious and bogus messages that disrupt the system’s normal operation. As a result, we required to verify the message sender’s identity. Due to its greater security over single-component user authentication, multi-component user authentication is becoming increasingly popular. Existing security and privacy authentication schemes, on the other hand, necessitate repeated interactions between users and various authentication components when sensing many components, resulting in additional costs and poor user experiences. Most of the existing schemes is not efficient to perform the certification validation due to the transmission overhead. As a result, these techniques are not suitable to implement real-life problems. Different types of authentication schemes are analyzed to perform the comparison between the varieties of components to improve the security requirements and efficiency in this paper.

Nowadays, data security has gained much importance. Encryption has evolved as a solution and is now a necessary component of every information security system. A variety of processes are necessary to protect the shared data. The present research focuses on employing encryption to protect data as it travels across the internet. This paper proposes that the data being sent from sender to receiver in the network must be encrypted using the AES-128 encryption approach. Decryption techniques allow the recipient to access the original data. This study examines the methods for manipulating messages and data. Most of the time, these attacks substitute encrypted data with corrupted data, resulting in unwanted data. These data are included by attackers throughout RTL design, physical design layout level, and maybe manufacture levels. Hardware Trojans are the term for them. However, there is a countermeasure for every assault, and therefore in this work, a deception of attacker approach is proposed to enable data transaction.

Advanced automotive electronic-control technology has led to added electronic equipment in the vehicle. Frequencies and power have gradually increased in the vehicle, creating a denser atmosphere of electromagnetic waves. The automotive components tend to produce Electromagnetic Interference (EMI). This will greatly contribute to EMI in the vehicle, thus disturbing electrical/electronic equipment and possibly damaging electrical/electronic components. Implantable Cardioverter Defibrillator (ICD) monitor your heart rate with a battery-powered device placed under the skin. ICD is connected to heart using thin wires. An electric shock will be delivered if it detects an abnormal heart rhythm if your heart is wildly beating and too fast. These medical devices tend to malfunction when they are exposed to high levels of EMI radiation. Electromagnetic compatibility (EMC) regulations must be met by the automotive industry and individual vehicle manufacturers. Regulatory compliance to EMC standards and CISPR 25 for automotive application circuits must be met which are not achieved under certain circumstances. For medical devices like ICDS CISPR 22 standards must be met. The radiation from EMI is capable of causing pacemaker or ICD malfunctions, ICDs to send shocks that aren’t needed, and EMI can interfere with PA sensor readings. Here, two types of filtering techniques have been designed in order to mitigate the CM noise which affects performance of Electromagnetic Compatibility (EMC). The results were compared, and the effective filtering solution is identified to reduce the conducted emission due to common-mode noise for the range of frequency from 150 kHz to 18 GHz according to the standard. As per EMC standard CISPR 22, the RF inductor is depicted and investigated within 150 kHz–30 MHz frequency range. The capacitive and parasitic impedance are calculated and used in the simulation process. This allows us to reduce the noise by as much as 30 dB which is an efficient noise reduction.

This paper presents two conformal antennas for wireless capsule endoscopic systems for robust biotelemetry communications, one of them covers 865 MHz band and the other one covers both 865 MHz and 2.4 GHz ISM bands. Biotelemetry is one of the applications of these bands. Two planar antennas are designed first and then they are conformed around a cylindrical capsule. The simulation is performed in ANSYS Electronics Desktop simulation software and it is observed that the antennas have wider bandwidth at respective bands. The peak realized gain for both antennas at 865 MHz are −18.1 dBi while antenna with bow and arrow shaped radiator has peak gain of −13.3 dBi at 2.4 GHz. Since the capsule needs to be ingested by the patient for endoscopy, biocompatibility has been ensured. The maximum specific SAR value is obtained for both the antennas and is considered safe in accordance with the IEEE standard safety guidelines. The recommended antennas’ conformal design idea, omnidirectional radiations, and multiband capability with wider bandwidth in the ISM band(s) will improve the scope for capsule endoscopy and would provide valuable contribution in the domain of biotelemetry.

Videos are composed of multiple tasks. Dense video captioning entails captioning of different events in the video. A textual description is generated based on visual, speech and audio cues from a video and then topic modeling is performed on the generated caption. Uncertainty modeling technique is applied for finding temporal event proposals where timestamps for each event in the video are produced and also uses Transformer which inputs multi-modal features to identify captions effectively and to make it more precise. Topic modeling tasks include highlighted keywords in the captions generated and topic generation i.e., category under which the whole caption belongs to. The proposed model generates a textual description based on the dynamic and static visual features and audio cues from a video and then topic modeling is performed on the generated caption.

Over the last few years, autonomous vehicles (AVs) have attracted researchers from different disciplines. Despite the effort put into the understanding and development of AVs, little is known about the factors that affect the public attitude towards this emerging technology. On the other side, public attitude is the main factor for the success of emerging technologies. Additionally, over the last few years, AVs were involved in multiple accidents that negatively affect the public attitude towards AVs. This paper focuses on investigating the impact of AVs’ accidents on the public attitude towards AVs for different groups with different demographic characteristics (age, gender, educational level, household, and income) in the USA by analyzing the opinions of 5880 respondents who participated in the online questionnaire survey. The results show that people become less positive towards AVs after accidents but with different magnitudes for respondents with different demographic characteristics. Additionally, the results shed light on the importance of educating the public about the benefits and state of technology of AVs.

This chapter investigates mathematical models of radar echo signals. The correlation properties of these signals were evaluated. A method of increasing the radar resolution based on the autocorrelation estimation algorithm is proposed. Potential possibilities of the method are investigated. The main contribution of this chapter is the consideration of the signal detection algorithm in accordance with the signal database. The main goal is to increase the degree of their resolution and increase noise immunity during coherent processing. A model of the radar echo signal is presented, taking into account the variety of possible probing factors. The experimental results made it possible to estimate the potential gain in terms of signal/noise when using the method. The resulting increase in resolution reached several times at a fixed signal/noise level. The findings will be useful in improving radar signal processing algorithms taking into account fluctuation components in their complex envelope.

Recent advancements in synthetic cell construction have made possible the begin of a research program whereby these man-made systems, which resemble biological cells at a minimal complexity level, can be conceived as tools for investigating information and communication theories in the “wetware” domain. In this paper, we will firstly present the field of synthetic biology and the features of synthetic cells (in particular, synthetic cells built from scratch). In the practical field, their potential role as “smart” drug delivery agents is probably one of the most ambitious goals, which needs a well-conceived SC design and advanced features. The latter includes sensing and perception, information transduction, control and programmability. These considerations elicit, at the same time, more general and theoretical questions, here presented as a sort of programmatic discussion. We ask whether and at what extent synthetic cells can be considered a valuable platform for investigating AI, cognition, communication, evolutionary optimization in novel versions: the chemical ones. We will not deal on what AI offers to synthetic biology, but on what synthetic biology offers to AI. By depicting some research paths, here, we intend to stimulate the bottom-up synthetic cells community to look toward such themes, to develop chemical AI in basic and applied sciences.

Research on deep learning (DL) to do detection of non-orthogonal multiple access (NOMA) and OFDM is presented in this paper. The successive interference cancelation (SIC) is generally fulfilled at the receiver in NOMA systems that decode multiple users in a successively. The detection accuracy is mostly based on the true detection of previous users due to the effects of error propagation. The NOMA receiver based on DL is described with deep neural network (DNN), which implements an estimation of channel and detection of signal together. The receiver has robust characteristics on the power allocation of the user is explicit from the simulation results. DNN is suitable for both linear channels and nonlinear channels, also the receiver is getting well on detection while the number of users is increasing. DL approximation obtains better achievement than a ML detection that ignores interference effects when the interference of the inter-symbol is intense.

Early detection of the progress of Diabetic Retinopathy (DR) in people with diabetes is essential to prevent vision loss. Our project is based on the diagnosis of DR which relies on the prediction of disease stage in the patient via the fundus image taken by a retinal camera. As symptoms, we took blood vessels and exudates as the segmentation target. For classification, we used the deep learning approach to train our model. Toward the end implement a user interface to facilitate the diagnosis of retinal images. The results of the DR stage classification distinguish between six stages: no DR, moderate, severe, and very severe non-proliferative diabetic retinopathy (NPDR), and then moderate and severe proliferative diabetic retinopathy (PDR). The evaluation of the proposed methods is performed by calculating the following parameters: sensitivity, specificity, and accuracy rates.

This paper presented a development of a system that can emulate a real electric motor in real time. This system, called digital twin or real-time digital simulator (RTDS), has critical value for many applications. The digital twin was developed in the field-programmable gate array (FPGA) structure, which is a high-speed digital system for real-time operation. The motor model was rearranged so that it can be run in real time on the FPGA. The developed digital twin was operated together with the real motor under the same conditions. Both the armature currents and motor speeds of the two motors were examined instantaneously on the same scope screen. Digital twin motor results were achieved to be nearly identical to real motor results. Especially, the digital twin showed dynamics similar to real motor dynamics with high accuracy as expected for reference signals with very fast transitions. The latency of the developed system was measured as about $$10\, \upmu \text {s}$$ . It can be said that this latency is quite enough to emulate electric motors.

Security applications in remote sensing missions have grown in the recent decade since the confidentiality of data carried from satellite to ground station can be hacked, and the space environment can damage satellite hardware. This study provides a unique architecture for secure Land Surface Temperature via radiation FPGA based on the AES algorithm. The T-Box method was used for the implementation of the AES algorithm. The testing results demonstrate that the suggested implementation achieved a throughput of 524.26 Mbps using Xilinx Virtex-4QV FPGA, at a maximum clock frequency of 208.453. Furthermore, the proposed solution occupies 8% of the slice in terms of area consumption. Moroever, an evaluation of the suggested method for prior implementation reveals that the presented approach has the best trade-off between area utilization, frequency, and throughput. Therefore, The suggested approach offers excellent performance, making it appropriate for future satellite remote sensing missions that utilize radiation-tolerant LST-SW computations.

The popularity of WiFi technology opens many new attack opportunities for attackers. It is a common practice to deploy an intrusion detection system to mitigate these attacks. In recent years, a few research studies have used different machine learning techniques to empower the intrusion detection system, hence improving the detection performance. However, most of the published methods do not consider the relationship between network traffic, so these methods consider the incoming traffic flows as independent traffic. In this paper, we employ graph neural networks to learn the relation between incoming network flow. The experimental results show that we can improve the performance of intrusion detection systems.

This paper presents the design and implementation of a power failure monitoring system for 22-kV transformer substations using the Internet of Things (IoT) technology and LoRa communication. The proposed system consists of 3 main components: Sensor nodes consist of voltage sensors and a LoRa transceiver module installed at the substation; Gateway station is composed of one LoRa transceiver module, one Wi-Fi module and one 3G/4G wireless module for transmitting data to Cloud server through internet connection and sending SMS alerts to users; and the ThingSpeak cloud server. The application of LoRa communication technology allows monitoring of grid voltage in real-time with lower operating costs than the methods of sending data over the internet and mobile phone networks. The research results can be applied to automate the monitoring and warning of power failures at the 22-kV distribution substations to help improve the stability of the power systems.

The properties of two different types of optocouplers, a conventional bipolar one with phototransistor output stage and a photorelay with power MOS output stage, have been determined before, during, and after irradiation with 68 meV protons with a fluence of up to 1e12 protons/cm2. In-situ measurements of the radiation-induced current of the input LEDs and in the case of the bipolar optocoupler also of the current-transfer-ratio of the device enabled a separate evaluation of input and output device degradation. A moderate degradation of the LED performance is observed for both devices, but in the case of the MOS-based photorelay, the most important overall device parameters are still within the specifications even for the highest irradiation level, while for the coupler with bipolar phototransistor a more than one order of magnitude decrease of the current-transfer-ratio is already observed for moderate fluences. This strong degradation is mainly due to the strong loss of photo-generated charge carriers with increasing fluence.

Communication is the most important process for a living organism. Autism is a complex developmental disorder that causes communication problems. The only way to communicate for babies is through crying, so screening analysis is an attractive approach to early diagnosis of autism to improve recovery. The aim of this paper is to explore the potential of machine learning (ML) in babies’ crying decoding toward autism diagnosis, pain identification, baby needs, etc. An extended literature review on the subject has been made to provide sensible comparisons between recently introduced ML approaches. Research findings reveal that ML can facilitate infant needs decoding and early diagnosis through crying pattern analysis; however, available infant crying datasets are scarce.

Cloud computing is becoming adopted increasingly by organizations all around the world. This is because the cloud is proven more dependable, affordable, elastic, and secure than traditional storage. However, this rapid and huge growth has opened the door for a great number of threats against the data in the cloud. Therefore, EaaS “Encryption as a service” was proposed to protect the data at all times. In this paper, we study, evaluate, and analyze the performance of three cryptographic algorithms, 3DES, AES, and Blowfish in terms of encryption time, decryption time, and memory consumption (Throughput). The aim is to provide more insights into the implementation of each and determine which provides more security with lower memory cost and execution time to be used in the EaaS solution that depends on crypto-algorithms to secure data and privacy in a cloud environment.

Energy exchange takes place between turbine and generator in the power system during subsynchronous resonance (SSR) which leads to torsional interaction between the shafts. Resonance in the power system is caused by the series capacitors connected to the transmission line. This paper aims to present an electromechanical approach to analyse and interpret subsynchronous resonance using the Finite element method. Subsynchronous resonance is introduced in two test rigs consisting of turbine, generator, shaft, and coupler with capacitors. Experiments and simulations (torque analysis and frequency response analysis) are conducted in test rigs and ANSYS workbench 16.0. Moreover, a spring damper is modelled to improve the stability of the shaft. From the results, it is clear that mechanical stress is increased when capacitors are connected to the test rig. A spring damper is installed at the point where the deformation is high. The damper reduced the stress and the vibration.

In the present work, the possibility of studying power electronic devices using platforms for modeling IP networks is considered. This research attempts to confirm or deny the hypothesis that power electronic devices generate very little traffic and connecting them to a working IP network would not lead to any change in the working performance of an IP network. Different scenarios with different topologies of modeled IP network are created to test the hypothesis. For the purpose of the study, the GNS3 IP network modeling platform and various IP network monitoring tools are used.

Employing individuals via the Internet has been a boon for businesses in the modern day. It is much simpler and more convenient than traditional recruitment methods. However, several scammers are abusing this platform, which may result in financial and privacy loss for job seekers and damage to the reputable organisation's name. In this research, we proposed a technique for detecting Online Recruitment Fraud (ORF). This model uses a publicly available dataset containing 17,780 job postings. We apply the four classification models to determine which classification model performs best for our suggested model. In this model, we use decision trees, random forests, Naive Bayes and logistic regression methods. We have estimated and evaluated the accuracy of several prediction systems. The random forest classifier provides the greatest accuracy, 97.16%, on our dataset. We have endeavoured to develop a method for detecting bogus recruiting postings.

Multiplier is one of the functional blocks in the Arithmetic and Logic Unit. Designing a delay-optimized multiplier is always a challenging task at the system design level. In this paper, a 4:3 counter is proposed based on 2-bit reordering circuit. The Wallace tree multiplier (WTM) is designed using 4:3 counter and other adder circuits such as 6:3 counter, Full Adder (FA), as well as Half Adder (HA). The designed WTM is utilized in the design of Urdhva Tiryagbhyam (UT) sutra-based multiplier. In addition, the complex multiplier is designed using the proposed Vedic multiplier (VM). The designed complex multiplier’s functionality is verified by Xilinx Vivado 2017.2 and also synthesized the circuits by targeting ‘xc7s50fgga-484-1’ device of Spartan-7 family. Furthermore, the performance of complex multiplier designed with proposed VM is compared based on the parameters such as Critical path delay (CPD) and LUT count with the existing multipliers.

In the recent era, data security is important for multimedia communication such as image and video. Secured communication and confidentiality of data play an important role in many online services like authentication, video conferencing, online classes, etc. In this paper, a video encryption technique using improved logistic map and DNA sequencing has been proposed to encrypt the video. The proposed video encryption technique has mainly three phases, firstly key is generated using SHA-256. Second, permutation of the video frame is completed using improved logistic map and SHA-256. In the third step, diffusion on permuted video frame is performed using SHA-256 key and DNA sequencing. Efficiency of the encryption technique is measured through PSNR, entropy, correlation coefficients and NPCR. The presented encryption technique is resistant to the brute force and statistical attack.

The objective of this work is to design, develop and demonstrate optimized fuzzy controllers for processes presenting nonlinearities such as coreless DC micromotors speed control problems, in order to improve the overall control system performance. One type-1 fuzzy logic controller (FLC) and one type-2 FLC (both Tagashi-Sugeno based) were designed and implemented in real time to improve motor speed control performance. For this purpose, a high-end microcontroller was used in which the developed controllers where embedded. The parameters of the input and the output membership functions of the proposed FLCs were optimally tuned using a well-known optimization algorithm, i.e., the particle swarm optimizer (PSO). A comparison of type-1 and type-2 Tagashi-Sugeno fuzzy control architectures is performed in different experimental scenarios and the relevant results are discussed and analyzed. The results demonstrate that type-2 FLC exhibits very satisfactory motor driving capability, providing zero overshoot, faster rise time and low settling time.

Leakage current consumes the majority of the active mode energy in the high-performance integrated circuits of today. In particular, in high microprocessor and system-on-chip architectures, the SRAM cell array is the primary source of leakage current. So low-leakage SRAM design is essential. Today’s VLSI designs are all about cutting down on power dissipation, supply voltage, leakage currents, and chip area. However, raising these parameters increases sub-threshold leakage currents and power dissipation, reducing design performance. Increasing the cell area reduces leakage power dissipation in standby mode. To solve these problems, it is best to cut down on effective leakage currents and dynamic power dissipation. For low voltage and energy constraints, the power dissipation, area, and delay performance of the low-power design of the 6T-SRAM cell with the DTMOS technique for a proposed low-power SRAM will be implemented in the Tanner EDA tool in 22 nm technology. The power of both types of SRAM cells will be compared using low-power techniques and a power analysis.

Recent years have seen an increase in instances of diabetes mellitus, a metabolic condition that if left untreated can severely decrease the quality of life, and even cause the death of those affected. Early diagnostics and treatment are vital for improving the outcome of treatment. This work proposes a novel artificial intelligence-based (AI) approach to diabetes classification. Due to the ability to process large amounts of data at a relatively quick rate with admirable performance, the XGBoost approach is used. However, despite many advantages, the large number of control parameters presented by this algorithm makes the process of tuning delicate and complex. To this end, the planet optimization algorithm (POA) is tasked with selecting the optimal XGBoost hyperparameters so as to achieve the best possible classification outcomes. In order to demonstrate the improvements achieved, a comparative analysis is given that presents the proposed approach alongside other contemporary algorithms addressing the same classification task. The attained results clearly demonstrate the superiority of the proposed approach.

A dental X-ray vividly exposes all of the features of the mouth that cannot be seen physically, such as hidden dental structures and bone loss. Cavities in teeth may be detected individually from X-ray pictures because a change in brightness correlates to a change in depth, surface orientation, material quality, and difference in scene illumination. The quality of a cavity differs from those of a healthy tooth. The X-rays must be processed using the appropriate image processing techniques. Most X-ray scans include noise initially, which degrades image quality. As a result, several image processing techniques are employed in the detection of dental cavities utilizing dental X-ray images. A literature review was undertaken on image enhancement, image segmentation, and feature extraction techniques for automating the procedure of dental cavity detection.

In the case of digital applications, addition is the most often utilized mathematical operation. Because they impact floating-point and arithmetic logic units, as well as cache/memory address computations, the stability of FA cells is considered to be critical. Full adders are critical elements in applications like DSP systems and microprocessors. The design of polar decoder-based full adders is crucial since the polar decoder architecture is extensively used in majority of the digital systems including processors. As a result, adder design is important in digital design. This study looks into Cadence implementations of polar decoder-based full adders in 180 nm and 90 nm technologies, with consideration of delays and power consumption.

Linear Feedback Shift Register (LFSR) is fundamentally a shift register capable of generating random sequences. It is a Pseudo-Random Number Generator (PRNG) whose randomness is driven through the linear feedback function governed by the primitive polynomial. LFSR has many real-time use cases but is not limited to cryptographic keys, NONCE, fast digital counters, data whitening, ATSC digital broadcasting standards, Intelsat business service, CDMA cellular telephony, Ethernet scrambles, etc. Due to the enormous growth in the VLSI industry, optimised LFSR designs were constructed to fit inside the chip’s silicon substrate to perform the indented tasks. The Complementary Metal Oxide Semiconductor (CMOS) technique is largely adopted over the globe for the design and implementation. This work investigates the effect of LFSR design over two different CMOS technologies, namely 90 nm and 180 nm, through the pre-layout and post-layout simulations. Further, the results have been compared among themselves. Fascinatingly, LFSR using 90 nm occupies a very less area footprint of 90.07 μm2 and 0.04371049 W of power consumption in a Cadence Virtuoso environment. The area occupancy of 90 nm LFSR is 29.43% less than the 180 nm LFSR, which shows the impact of technology mapping. In addition, the proposed LFSR design is compared with the existing LFSR designs of various technologies, and the results ensure the consistency of the CMOS-based LFSR.

Image is well-known word in various fields like medical, engineering, and arts. A literature review is conducted on image qualities, and the need to improve the quality of image for various applications are identified. Numerous noises are added to an image both internally and externally in different stages which results in the poor quality of image. At few points, the noise in the image leads to loss of critical information and creates a lot of damage in that area. Purpose of this review is to gain knowledge about the image denoising and give details about denoising filter algorithms used for noise removal from images. Various techniques exist to overcome noise in any image. Multiple filters and combinations of filters have been designed to remove noise, and those are summarised in this paper for better understanding.

Billions of images are uploaded daily, and it requires a large storage space. Utilization of better storage capacity and to improve uploading/downloading time, researchers have designed an image compression model. Many researchers have implemented various approaches to improve the image compression ratio of an image. This paper presents an analysis of various optimization algorithms based on vector quantization (VQ). The first algorithm is a modified genetic algorithm. It is based on Darwin’s principle which is natural characteristics. Those who are fit can survive and use it to optimize the codebook. A second algorithm for optimization of the codebook is particle swarm optimization (PSO). PSO algorithm is superior to finding the codeword vectors of codebook from the training image samples for image compression. In the PSO algorithm, the selection approach plays an important role to select the particle based on the fitness of the population. Training images from the standard image database are used for the design of the codebook. The input image set is 4 × 4 or 8 × 8 blocks and is represented as vectors. They are referred to as codewords in the codebook, and it is a component of a code. The codebook size is measured by codewords. The block size is decided by the length of the codeword. These codewords generate the codebook by entering the vector value. Compression is done with the help of sending indices to the decoder. Likewise, analysis of quality measures is presented to the modified GA and PSO algorithms based on mean square error, peak signal-to-noise ratio, structural similarity index, and average difference. In this work, we have calculated bits per pixel (BPP), the compression ratio (CR), and the % compression ratio. The experimental results are validated.

This paper represents a comparative simulation study of I–V characteristics of GaAs/InP and Si/SiC heterojunction solar cells. The design and simulation of device is done with COMSOL at 300 K. The calculated fill-factor of GaAs/InP and Si/Sic is 0.75 and 0.85, respectively. Moreover, open-circuit voltage (Voc) of the GaAs/InP heterojunction solar cell is 0.4 V with efficiency of approx. 10%, whereas that of Si/SiC is 0.52 V with efficiency of approx. 14%. The effect of the material properties on the device with I–V and J–V curve is very well demonstrated.

The data exploration metrics through scientific mapping aim to find out the level of interest in social networks and politics 2.0 within the scientific community worldwide; the development of literary production in times of pandemic to which research is limited; the creation of both research and review articles sectioned by countries, authors, and keywords within the search carried out in the Scopus database. In the first instance, an analysis of the scientific literature is carried out in the primary citation index that is used worldwide to measure high-impact scientific production, such as Scopus; a total of 139,710 jobs were obtained, which were filtered only in the years of the COVID-19 pandemic, resulting in several jobs of 17,049. The results that could be analyzed through scientific mapping showed that scientific production increased in the keyword social networks (online) with several authors who mostly belong to developed and English-speaking countries; Mexico is the only country in Latin America within the results obtained through the maps. Finally, the analysis of the clusters relates keywords such as social networks, machine learning, and big data; within the systematic review, there are still no studies related to politics 2.0; therefore, it is concluded that the political keyword 2.0 still has no connection in terms of high-impact scientific production.

The use of technology has invaded unprecedented professional areas that we could never have imagined, such as medicine, engineering, among others. The employ of robots has spread throughout the world mainly in manufacturing companies where they have reduced production times and costs and are guaranteeing greater production and quality in products. The objective of this research is to identify new challenges for organizations to visualize the harmony between technological growth and human talent. In addition, it is showing that the arrival of the new industrial revolution brings with it dramatic changes in job profiles. This research was carried out using data mining with the development of decision trees that allowed us to address the current problems of artificial intelligence and the replacement of human talent as a higher education case study. Applying this prediction and data segmentation technique, information was obtained that was analyzed for future decision-making in the field of education and the application of artificial intelligence in the study population. In addition, a structured questionnaire validated by experts was designed for data collection and reliability was measured with Cronbach’s Alpha coefficient. Subsequently, the instrument was applied to teachers and students of higher education in the province of Tungurahua, with 100 informant agents. As a result, companies must seek a balance between machine and man. In addition to replacing humans, robots could be incorporated responsibly into work environments.

It is very challenging to predict a crime scene only by machine without human intervention. This research has tried to make that possible. Convolutional neural network (CNN) has been used to detect 4 objects which are handgun, fire, knife, and accidents. By detecting these objects easily with the help of CCTV cameras, the machine can predict the crime scene. Machines will be able to identify crimes swiftly and intervene based on situations like accidents or violence. This research has adopted a variety of techniques to reach the pinnacle of implementation and success. The model used here has been built with the help of CNN, and there are 4 objects to classify which are mentioned earlier. This research has succeeded in predicting crime scenes through CCTV cameras which may bring prosperity to the country and the nation.

Physical parameter monitors are necessary for the conversion of conventional factories to smart factories. The creation of new sensors for monitoring physical parameters in difficult-to-access areas is necessary for this move. In this line, a large number of optical sensors based on integrated optical wave guides or optical fibers have been designed and manufactured over the previous ten years. As a potent tool for real-time monitoring of physical parameters like temperature, pressure, strain, and humidity. Fiber Bragg Grating (FBG)-based sensors have attracted a lot of attention. The main reasons for using FBG sensors in smart factories are immunity to electromagnetic interference and radio frequency; compact in size and offer multiple sensing to different physical parameters; permit remote sensing and are not prone to corrosion. In this paper, a review of FBG-based sensors for strain parameter monitoring and their application in the smart factories is presented. An overview of the historical background is followed by an explanation of the fundamentals of FBG sensing and a discussion of the electromagnetic theory of waveguide modes in optical fibers. Then give a review of current research in FBG strain sensor development. A review of the challenges and applications of FBG strain sensors specifically in smart manufacturing follows.

Recently, the commercialization of various convenience services using beacon on the Bluetooth low energy technology is on the rise as one of the results of technological innovation by the information communication technology, which rapidly develops with the coming of the fourth industrial revolution. In addition, due to expansion of the scope of information communication infrastructure and things such as “smart city”, a large-scale convergence contents platform based on the IoT becomes one of the major issues in the IT industry. However, the technical measure concerning the management and processing of data collected by wireless sensor devices such as beacons is still unclear, and the concrete solutions to this problem are currently needed. Therefore, this study suggests one basic point about the technique to analyze and manage the large data collected by the beacon. It also efficiently provide the converged contents service using the user information data stored in the server. The development of this application enhances the social utilization value of beacon service in the future.

The road accidents caused by driving of young and inexperienced drivers have been increasing day by day. Due to the rise in catastrophic accidents, there have been several serious injuries and damages. Therefore, age detection utilizing deep learning is established here in order to alert and stop these massive calamities. In this system, the CNN model is used to predict the age from a face image, and the age classifications task is developed as a classification problem. By predicting the age from the faces and barring the children from driving, the proposed method aids in keeping our children safe. Here first detect faces from the input video stream, then extract the face Region of Interest. Then, in order to estimate the age of a person, an age detector algorithm is employed here. In this system for face detection, Open CV’s DNN module and DNN face detector are used. In this proposed system, the face is detected from the real-time video frames via a web camera. Then predict the age of the person from the face. If minor aged person is detected, then car cannot be started by that person and the system alert the guardians by alarm. If the detected person’s age is above underage, then the car will start automatically. The system will also send an alerting SMS to the guardians using GSM module. The proposed system is intended to identify underage drivers and stop them from operating a vehicle.

The duty cycle MAC technique has been deployed in wireless sensor networks to decrease sensor node energy utilization. EC-MAC, or Energy-conservation Medium Access Control, is used to reduce energy consumption. The primary goal is to improve sleep latency while also balancing energy usage across sensor nodes. The suggested EC-MAC is compared to the RI-MAC and the outcome is shown. The simulation is done in NS-2. The presented method produces better results in terms of packet delivery ratio, overall energy, duty cycle, and leftover energy, according to the experimental results.

According to the American Cancer Society, cancers related to the brain and nervous system are ranked as the tenth leading cause of mortality in humans. In addition to this, the World Health Organization (WHO) reports that low-income nations are experiencing a lack of neurologists, who play an essential part in the functioning of the healthcare sector. There is currently no method that is reliable enough to permit the classification of brain illnesses into multiple classes. The multi-class classification of clinical brain images was made possible by our machine learning approach, which we proposed. The classification of brain disorders, including Alzheimer’s disease, dementia, brain cancer, epilepsy, stroke, and Parkinson’s disease, would be accomplished using a deep learning-based convolutional neural network (CNN). The Visual Geometry Group-16 (VGG-16) architecture was taken into consideration throughout the feature selection process, and the Adam optimizer was used to perfect the model. The proposed CNN model would be beneficial in alleviating the arduous labor of neurologists.

One of the fatal diseases in the world is heart disease. Every year, millions of people die of cardiovascular diseases. However, one can decrease the mortality rates if the heart disease was detected and treated early. Usually, people do an electrocardiogram (ECG) test to know about the well-being of their heart. Some kind of irregular functioning and illness in the heart can be found in an ECG test. When the heart malfunctions or if there is any improper beating of the heart, then it results in arrhythmia. There are several types of arrhythmia and some of them are fatal. The process to identify the correct type of arrhythmia is quite difficult and effort-taking process. Even the small changes in the ECG relate to another kind of arrhythmia. It takes experience and patience to recognize the type of arrhythmia accurately. Therefore, deep learning techniques should be employed to analyze the test. Machine learning that involves many levels of processing is known as deep learning. From computer vision to natural language processing, there’s a lot to learn. It has been used in various applications. This method is receiving more popularity because of extreme accuracy, provided the numerous amount of data. The interesting feature is that it analyses the examples and distinguishes the classes and levels automatically. This study is regarding arrhythmia prediction in ECG and the attention it deserves in deep learning community. Providing CNN model, we are going to elaborate the process of detecting cardiac arrhythmia using ECG dataset in this study. The model is executed by rendering CNN with cardiac arrhythmia recognition database. Purpose: About one-third of the world’s population is affected by arrhythmia. Hence, the development of new and successful methodologies is highly in demand in the field of arrhythmia prediction. Further, the need of a cost-effective wearable monitoring gadget to identify the condition of arrhythmia is highly recommended. It assures the trouble-free environment for those who are affected. Observations: Various research papers that were written bases on arrhythmia prediction using machine learning techniques. Additionally, there are also new advancements all over Internet regarding deep learning-based strategies. These strategies can bring an immense change in cardiac arrhythmia prediction.

The society’s convenience food intake practice has made culinary science one of the blazed areas of food industries. Higher quality food standards and convenience require good knowledge on culinary science. Unlike chefs, consumers are not much aware of the culinary skills. Moreover, working adults are coped with time pressures, often forget to supervise food under process. This leads to time and fuel wastage. This paper involves development of gas stove automation system that controls flame intensity to allow cooking under pressure cooking, boiling, and hybrid modes. The model employs parameters like whistle count, food surface temperature and cooking time, fed through touch display. The temperature module evaluates food surface temperature using a contactless sensor while employing random forest regression model. The gas knob is adjusted using stepper motor. The proposed model is expected to reduce lead cooking time and fuel consumption, improve dish quality and prevents any possible fire accidents.

Understanding the cognitive factors that contribute to introductory programming students’ abilities to learn to program is critical to helping computer educators create better opportunities for students to improve their programming performance. The goal of this research is to explore cognitive factors that have an influence on programming performance in introductory programming courses in particular. The study documents 17 factors from 25 empirical studies that analyzed the influence of these factors on programming performance. Our analysis shows a wide range of cognitive factors studied and interrelated groups of factors studied in literature focused on introductory programming courses. This is a valuable review of information regarding influencing cognitive factors to restructure aspects of future introductory programming course curricula to benefit students’ ability to learn to program.

When it comes to circuit design, digital elements have a pivotal and crucial impact. Input values given to these components can be degraded by internal and external disturbances, it can happen either by virtue of foreign factors, to the lack of efficiency in the sensors, or maybe due to unknown lags in the communication systems. Hence, it is vital when the output responses related to these devices stay as robust when slightest fluctuations occur. Layout of robust components has been reported as primary challenge in the implementation of electronic systems. Ways to improve the robustness of circuits have been dealt with but either by adding huge amount of extra logic, alter the circuit latency, or these are suitable only for such circuits like microprocessors. Also, they suffer limitations by capturing application particular information of the circuit. However, the proposed methodology requires only a small increment in the extra hardware, which only affects the timing characteristics of the circuit in a less considerable manner, and will spontaneously apply to any of the circuits which are arbitrary.

In recent decades, there have been several key events in the field of access to information such as the emergence of open and linked data repositories, the expansion of social networks and the development of mobile and web technologies. In this way, the necessary infrastructure has been created to be able to access and exploit the information immediately and in real time. In this sense, the use of applications that implement value-added services processing information from repositories and networks has been extended. This article describes a system that combines a mobile application and a web application that exploit information retrieved from Twitter, the open data repository of the Madrid City Council and an online newspaper, with the aim of implementing an alert generation service categorized in real time on events that have occurred in the city of Madrid.

Food waste is one of the biggest costs of concern in all over the country. Research papers roughly say food wastage is about 1,300,000,000 tons. This paper provides a detailed view of predictive analytics of ML together comparing different strategies to separate different types of food. Non-destructive food quality is checked by based on the outer appearance of the given sample without destructing and sensed by the given sensors. This proposed system employed with PIC microcontroller which acts as a central processing unit, hygrometer, and temperature sensors together with three types of gas sensors which send the data sent to the cloud. For classifying the different types of foods like vegetables, fruits, and dairy qualities, the data is sensed with different types of grove gas sensors MQ9 (CO, Coal, Gas, Liquid Gas), MQ3 (Alcohol Vapor), and MQ2 (Combustible Gas, Smoke) furthermore with environmental sensors were acquired and sent that data to the Internet of Things. The novelty of this system is used to predict the quality of food under climatic conditions and also in traveling time with the help of IoT and their android application and the estimation of time that can preserve the different types of food in the storage.