Journal of Science Engineering Technology and Management Science

The rapid proliferation of electric vehicles has necessitated the development of resilient and intelligent charging infrastructures capable of accommodating fluctuating demand whilst preserving grid integrity. In this regard, hybrid energy storage architectures, particularly those combining battery energy storage with fuel-cell systems, have emerged as a promising solution for grid-interactive charging stations. This paper presents a decentralised power coordination framework for an electric vehicle charging facility integrated with photovoltaic generation, battery storage, and a proton exchange membrane fuel cell. A modified line-resistance-compensated droop mechanism is employed to regulate power sharing among distributed sources, thereby ensuring stable direct-current bus operation under both grid-connected and autonomous conditions. Particular emphasis is placed upon state-of-charge-oriented battery regulation to restrain excessive depth of discharge, while the fuel cell is constrained to operate within its linear efficiency region to enhance durability and operational economy. The proposed strategy facilitates unity power factor interaction with the utility grid, sustains acceptable harmonic performance during islanding events, and accommodates dynamic charging scenarios without reliance on centralised communication. The effectiveness of the control methodology is substantiated through comprehensive simulation and experimental investigations, demonstrating improved power-sharing accuracy, prolonged storage life, and uninterrupted charging operation under diverse operating contingencies.

Keywords : Electric vehicle charging station; Decentralised power management; Hybrid energy storage; Fuel-cell systems; Battery state of charge; DC microgrids; Droop control strategy

Author : Mr. J. KANTHAIAH,Mr. K. RAMAKRISHNA PRASAD,Chimmana Naga Sanjay,Esam Rakesh

Title : A DECENTRALISED HYBRID FUEL-CELL AND BATTERY ASSISTED POWER COORDINATION SCHEME FOR GRID-INTERACTIVE ELECTRIC VEHICLE CHARGING FACILITIES

Volume/Issue : 2026;03(02)

Page No : 1-9

This project focuses on automatic video generation using artificial intelligence to simplify the process of video creation. Traditional video production requires significant time, cost, and technical expertise, which makes it difficult for students and beginners. The proposed system converts natural language text into complete videos by generating scripts, voice narration using Text-to-Speech, captions, and relevant background visuals. Natural Language Processing is used to analyze the input text, while video processing techniques combine all components into a final video. The system works through a user-friendly web interface where users only need to enter text, without requiring any video editing skills. It reduces manual effort, saves time, minimizes cost, and is useful for students, educators, and content creators, demonstrating how AI can simplify multimedia content creation.

Keywords : Artificial Intelligence, Text-to-Video Generation, Natural Language Processing, Text-to-Speech, Automated Video Creation, Multimedia Content Systems

Author : GOWTHAM G,ANAS MOHAMMED M,SARAVANAVEL V,SATHISHKUMAR J

Title : AN ADVANCED ARTIFICIAL INTELLIGENCE FRAMEWORK FOR GENERATING VIDEOS FROM NATURAL LANGUAGE TEXT

Volume/Issue : 2026;03(02)

Page No : 10-21

The quick development of smart home technologies and the growing interest in sustainable energy solutions, have pushed to merge renewable micro-generation systems on IoT energy management platforms. This paper shows the smart house energy management system design and its implementation based on a free turbine turbulence in micro system. The device captures air flow and turns it into energy - a task that the IoT-based microcontrollers and sensors monitor, control and manage. Data on turbine output, household consumption and environmental parameters are streamed using lightweight communication protocols to a cloud hosted platform. It offers a dynamic load balancing control between renewable generation, grid supply and optional energy storage for optimizing household energy consumption. A homeowners-dashboard platform delivers energy production and consumption data; a breakdown of savings and enables the ability to remotely turn on and off appliances. The system also includes predictive analytics to improve efficiency by predicting demand and distributing energy as needed. The IoT-based smart energy management system proves the viability of the micro-level usage of renewable generation and intelligent control with scalability and modern living sustainability concept. Employing a turbo wheel micro-turbine demonstrates the abilities of small renewable power systems in minimizing grid reliance and creating sustainable lifestyles.

Keywords : Smart Home, Energy Management System, Turbo Wheel Turbine, Renewable Energy, Microcontroller (Arduino).

Author : V.MALLIKARJUNAREDDY, Mr.K.SAIMAN, A.V.BHARATHREDDY, CH. SASI, M.MOHITSAI

Title : SMART HOME ENERGY MANAGEMENT SYSTEMS USING A TURBO WHEEL MICRO TURBINE

Volume/Issue : 2026;03(02)

Page No : 22-28

Brain tumors are a serious health problem, and detecting them early is very important for successful treatment. Normally, doctors confirm tumors through surgery and biopsy, which are invasive methods. In this project, artificial intelligence (AI) is used to provide a non-invasive way of diagnosis. By using MRI brain scans, deep learning models are developed to classify gliomas, meningiomas, pituitary tumors, and healthy brain tissues. The main model used is a Convolutional Neural Network (CNN), which learns patterns from a large dataset of MRI images. To improve accuracy, preprocessing techniques such as normalization, scaling, and data augmentation are applied. These steps make the system stronger and reduce false positives and false negatives. The proposed model shows high accuracy in separating tumor and non-tumor cases, helping doctors make faster and more reliable decisions. This project demonstrates how AI can support medical professionals, reduce manual effort, and provide precise forecasts. Overall, it highlights the importance of automated systems in medical imaging, making diagnosis more efficient, trustworthy, and useful for early detection of brain tumors.

Keywords : Brain Tumor Identification , Deep Learning with CNNs , Healthcare Image Processing , Intelligent Diagnosis Systems , AI Applications in Healthcare

Author : SK.FAZARUNNISA BEGUM, Mrs. K. RAJYAM , Y. AMULYA , CH. HARSHA VARDHINI, G. ANUSHA, T. AMULYA

Title : AUTOMATED BRAIN TUMOR DETECTION USING IMAGE PROCESSING TECHNIQUE

Volume/Issue : 2026;03(02)

Page No : 29-36

The requirements for portability have required rapid development of low power electronic communication products. The variable gain amplifier (VGA) is one of the weakest components of any modern receiver in terms of a correct signal baseband processing. Most CMOS VGAs have low bandwidth and limited gain range. In this paper, two stage class AB VGA consisting of direct transconductance amplifier and linear transimpedance amplifier is designed in Silterra 0.13-µm CMOS technology using Mentor Graphics environment. According to the post-layout simulation findings, the VGA design achieves the highest gain range of -33 to 32 dB and the broadest bandwidth of >200 MHz. The VGA merely uses 2 milliwatts of power. a single DC 1.2 V source. The core chip area of the VGA is likewise only 0.026 mm2 which is also the lowest compared to current researches. For all contemporary communication gadgets, such a VGA will be a very helpful module.

Keywords : CMOS, Receiver, RF, VGA

Author : G. VENKATA BHAVANI PRASAD, Mr. P. SIVA KRISHNA, K. GOPI KRISHNA, CH.VENKATA SRINIVASA RAO, S. VENKATA SAI CHARAN TEJA, D. PAVAN KUMAR

Title : DESIGN OF A LOW-POWER COMPACT CMOS VARIABLE GAIN AMPLIFIER FOR MODERN RF RECEIVERS

Volume/Issue : 2026;03(02)

Page No : 37-45

Imaging methods like Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), and Single-Photon Emission Computed Tomography (SPECT) have given medical professionals information about the soft tissues, structural features, and other aspects of the human body. distinct sensors collect distinct image information of the same part, and different imaging techniques maintain different properties. Better contrast, fusion quality, and perceived experience are the goals of the fusion. The outcome of the fusion should satisfy the following requirements: (a) the fused image should fully preserve the information of the source images; (b) the fused image should not generate any artificial information, such as artifacts; and (c) poor states, such as noise and misregistration, should be avoided. The suggested network is one of three CNN models used to compare patch similarity. Because its two weight branches are the same, the source image's feature extraction and activity level measuring methodologies are also same. This has some advantages over pseudo and 2- channel models, and the siamese model is also used in fusion applications due to its ease of training. After getting the weight map, the Gaussian pyramid decomposition is used, followed by the pyramid transform for multiscale decomposition, bringing the fusion process closer to human visual perception. Additionally, the decomposed coefficients are adaptively adjusted using the localized similarity-based fusion technique. The algorithm combines the popular pyramid-based and similarity-based fusion algorithms with the CNN model to provide a better fusion approach

Keywords :

Author : A.Yohan, Mr.P. Surya, M.Venkatesh, K.Nithin Reddy, K.Bala Adharsh,Y.Jagan Mohan Reddy

Title : DEEP LEARNING-DRIVEN MULTI-MODAL MEDICAL IMAGE FUSION USING CNN FOR ENHANCED BRAIN TUMOR DIAGNOSIS

Volume/Issue : 2026;03(02)

Page No : 46-51

A vital prerequisite for healthcare services is ensuring blood safety detection prior to transfusion. Transfusion hazards might arise from the presence of contaminants, clots, or aberrant aggregation in blood samples, even when blood group compatibility is proven. The manual microscopic observation used in conventional inspection techniques is labour-intensive. This paper presents an automated method for analysing transfusion compatibility using image processing techniques. The suggested method uses segmentation, noise filtering, grayscale conversion, and Canny edge recognition to handle microscopic blood pictures. To find anomalous structure patterns, morphological operations and numerical edge feature extraction are employed. The system decides if the blood sample is appropriate for transfusion based on the extracted attributes. For medical laboratories, this approach offers a quicker and more dependable initial screening tool while minimizing human participation.

Keywords : Blood Sample Analysis, Image Processing, Edge Detection, Medical Imaging, MATLAB, Automation

Author : Y. PRIYANKA, V.JAI KUMAR, V. KANCHANA, D. POOJITHA, M.V. MANASA

Title : BLOOD TRANSFUSION SUITABILITY DETECTION USING IMAGE PROCESSING

Volume/Issue : 2026;03(02)

Page No : 52-58

In this study, we created a unique modular neural network for precise human lung cancer diagnosis. we have taken the MRI reports of the patients and then we have evaluated the MRI images with the use of image processing techniques and neural networks to check whether the patient has been impacted by lung cancer or not. In order to improvise the MRI images for analysis we are grayscale function for making the images fit for the lung cancer. In order to determine the contrast and energy of the image, which are crucial factors in identifying lung cancer, we have applied the neural fuzzy classifications method. We have used the feature extraction approach to determine the image's entropy. We can determine whether or not the patient has lung cancer by getting the values of entropy, contrast, and energy.

Keywords : X-Ray CT Scan MRI Images, Artificial neural networks (ANN),Machine Learning, Lung cancer Detection, Medical Report Analysis

Author : P. GOVARDHAN, P. SIVA KRISHNA, P. PAVAN SRINIVAS, T. JAGADEESH, K. MANIDEEP

Title : DETECTION OF LUNG CANCER BY USING ARTIFICIAL NEURAL NETWORKS

Volume/Issue : 2026;03(02)

Page No : 59-64

Controlling electronic appliances in our homes is traditionally done through manual switches; however, with advancements in Internet of Things (IoT) technology, it is now possible to operate home appliances remotely from anywhere in the world. This project presents a smart home automation system integrated with a fish feeding mechanism that allows users to control and monitor household devices through a smartphone application. The system enables seamless operation of appliances with a single touch, providing convenience, flexibility, and enhanced control. The proposed solution is particularly beneficial for elderly and physically disabled individuals who may face difficulties in moving around and operating household devices manually. People living alone can also greatly benefit from such automation, as it reduces dependency on external assistance. In busy lifestyles, providing constant support to such individuals becomes challenging. Moreover, attempting to operate appliances manually may increase the risk of falls or injuries, which could lead to serious health consequences. By automating home appliances and essential tasks like fish feeding, the system minimizes physical effort, enhances safety, and improves overall quality of life through smart and efficient home management.

Keywords :

Author : Mr. A. GOPI, Mr. J. KANTHAIAH, K. GOPI CHAND,P. SRILAXMI, K. NARASIMHA RAO, P.SAI DEEPAK

Title : VOICE CONTROLLED HOME AUTOMATION USING MOBILE

Volume/Issue : 2026;03(02)

Page No : 65-72

This paper proposes a novel 10-transistor (10T) static random-access memory (SRAM) cell incorporating a single-ended, decoupled read bit line (RBL) supported by a dedicated 4-transistor read port to achieve lowpower operation and significant leakage reduction. The RBL is precharged to half of the supply voltage (VDD/2) and subsequently allowed to either charge or discharge depending on the stored data value. During read operation, an inverter controlled by the complementary storage node (QB) connects the RBL to dynamically controlled virtual power rails through a transmission gate. For a stored logic ‘1’, the RBL charges toward VDD, whereas for a stored logic ‘0’, it discharges toward ground. In both hold and write modes, the virtual rails are maintained at the same potential as the RBL precharge level and are connected to the actual supply rails only during read access. This dynamic control of virtual rails effectively suppresses RBL leakage current. Implemented in a commercial 65 nm CMOS technology, the proposed 10T cell occupies 2.47 times the area of a conventional 6T SRAM cell with ? = 2. It achieves a 2.3× improvement in read static noise margin (RSNM) and reduces read power dissipation by 50% compared to the 6T design. Furthermore, RBL leakage current is reduced by more than three orders of magnitude, and the ION/IOFF ratio is significantly enhanced relative to conventional bit-line leakage. Despite the additional transistors, the overall leakage characteristics remain comparable to the 6T cell while delivering competitive performance and improved robustness.

Keywords : T10T SRAM, Single-Ended Read Bit Line (RBL), Virtual Power Rails, Leakage Reduction, Low Power Design, Charge Recycling, Read Static Noise Margin (RSNM), 65 nm CMOS Technology

Author : K. Devaki Devi, M. Anusha, S. Ruchitha, K. Mounika, G. Gowthami

Title : DESIGN AND DELIBERATION OF A TEN-TRANSISTOR SRAM CELL EMPLOYING HALF-VDD ROW-WISE DYNAMIC PRECHARGE FOR DIMINISHED SWITCHING POWER AND ULTRALOW READ BIT-LINE LEAKAGE

Volume/Issue : 2026;03(02)

Page No : 73-80

This paper presents the design of a voltage-resilient Static Adaptive Energy (SAE) keeper-based full adder implemented using CMOS technology to achieve high speed, low power consumption, and robust operation under scaled supply voltages. The proposed architecture employs a keeper-assisted XOR–XNOR cell to overcome the critical challenge of maintaining full-swing output voltage in lowvoltage environments. By incorporating an optimized grounded keeper configuration derived from conventional 20-transistor structures, the design significantly reduces transistor count while enhancing performance metrics. The proposed 16-transistor full adder is simulated using TSPICE in 90-nm CMOS technology and evaluated across a wide supply voltage range from 0.6 V to 1.5 V. Performance comparison with existing keeper-based designs is carried out in terms of power dissipation, propagation delay, and power-delay product (PDP). Simulation results demonstrate that the proposed architecture achieves lower power consumption, reduced delay, and improved PDP across all operating voltages, confirming superior voltage resilience. The effectiveness of the design is further validated through implementation in a 4-bit ripple carry adder, making it suitable for lowpower and high-speed VLSI applications including arithmetic logic units, multipliers, calculators, and graphical processing units.

Keywords : Full adder-Hybrid logic design-Keeper circuit-Logical effort-Voltage resilient -Ripple carry adder

Author : G. Rajendra, Rayala Leela Naga Sai, Paladugu Gagan, Dhudalauday Sankar, Davireddi Sai Varma

Title : DESIGN OF A VOLTAGE-RESILIENT KEEPER-BASED FULL ADDER USING CMOS TECHNOLOGY

Volume/Issue : 2026;03(02)

Page No : 81-95

In the present age of digital intercourse and distributed computation, the preservation of secrecy and integrity in transmitted data hath assumed paramount importance. This discourse presenteth an FPGAbased high-security data transmission architecture founded upon the 128-bit Advanced Encryption Standard, yet contrived without reliance upon the traditional substitution box mechanism. The proposed design seeketh to mitigate the hardware overhead and latency commonly associated with SBox implementation, whilst preserving the cryptographic robustness requisite for secure communication. By reformulating the substitution transformation through composite field arithmetic and logical restructuring, the system accomplisheth efficient encryption and decryption suited to reconfigurable platforms. The architecture is realised upon a field-programmable gate array so as to attain parallelism, swiftness of operation, and adaptability. Careful optimisation of resource allocation and pipelining ensureth diminished power dissipation and augmented throughput. The resulting framework demonstrateth that formidable security may be achieved without extravagant hardware expenditure. Thus, the proposed scheme uniteth efficiency, adaptability, and fortified protection, rendering it apt for contemporary secure digital transmission systems.

Keywords : FPGA implementation, AES-128 encryption, S-Box elimination, Secure data transmission, Hardware optimisation, Cryptographic architecture, Parallel processing

Author : B.V.V.N.S.Sri Sowjanya, Allampalli Naga Phanindra, Bethala Madhu Babu ,Swathantra Pratap Singh, Bandaru Likhit

Title : AN FPGA-ORIENTED SECURE DATA CONVEYANCE ARCHITECTURE EMPLOYING A 128-BIT AES CIPHER WITHOUT CONVENTIONAL SUBSTITUTION BOX REALISATION

Volume/Issue : 2026;03(02)

Page No : 96-105

This paper explores the dynamic interplay between money and human relationships. From familial bonds to romantic connections and friendships, the role of money often shapes, influences, and sometimes disrupts the emotional fabric of human interactions. Drawing from psychological, sociological, and economic perspectives, this study delves into how monetary concerns impact trust, power dynamics, communication, and relational stability. The paper aims to provide a holistic understanding of how financial issues can both unite and divide people, with recommendations for fostering healthier financial dialogues and expectations within interpersonal relationships.

Keywords : Money, human relationships, financial conflict, trust, power dynamics, relational health, emotional capital

Author : Dr. Samir Kumar Naskar

Title : MONEY AND HUMAN RELATIONSHIPS: NAVIGATING VALUE, POWER, AND EMOTION IN MODERN SOCIETY

Volume/Issue : 2026;03(02)

Page No : 106-112

The swift growth of digital technology and internet uptake has had a huge impact on the retail sector with the adoption of e-commerce platforms being facilitated. When doing this research paper I have not only considered the dynamic aspect of online shopping but also I have examined the consumer behavior where e-commerce is concerned. The study will identify the key contributing factors of convenience, price, variety of products, reliance, web user experience and web payment systems on online purchase buying decision making by the consumers. It also investigates how demographic variables such as age, income, education and occupation influence the preference of online shopping. The research is based on the secondary and primary data. Primary data will be obtained via structured questionnaires which will be given to a wide array of online consumers whereas secondary data will be accessed via journals, reports and other reliable online sources. Data will be interpreted using such statistical tools as percentage analysis, means scores, hypothesis testing to reveal the trends in consumer behavior. The results demonstrate that the most influential factors affecting the adoption of online shopping include convenience, time-saving as well as appealing discounts. Also, the availability of secure methods of payment and easy-to-use policies of returns and positive reviews of customers increase their level of trust and contentment. Yet, the issue of data privacy, authenticity of products, and delayed deliveries are still a matter of concern in making decisions to purchase. Another issue that is raised in the paper is the growing role of mobile commerce and social media sites in the determination of consumer preferences and purchase intentions. It concludes the paper that online shopping has become a component of consumer experience these days and that younger consumers tend towards the online platform. It suggests that organisations ought to invest greater in user experience enhancing, high-degree level of cybersecurity protection, and individualised marketing practices in order to be competitive. To marketers, policy makers and researchers, the study is ingenious in the realization that the actions of the online consumer may be dynamic and future development of e-commerce sector may be achieved.

Keywords : E-commerce, Consumer Behavior, Online Shopping, Digital Marketing, Purchase Decision, Customer Satisfaction, Mobile Commerce

Author : Dr Rajat P.Ghorpade

Title : E-COMMERCE AND CONSUMER BEHAVIOR: A STUDY OF ONLINE SHOPPING TRENDS

Volume/Issue : 2026;03(02)

Page No : 113-123

The rapid growth of e-commerce and online review platforms has increased the prevalence of deceptive and fake reviews, which can significantly influence consumer purchasing decisions and undermine trust in online systems. Traditional fake review detection methods often struggle to capture the complex linguistic and behavioral patterns associated with fraudulent reviews. To address this challenge, this paper presents a hybrid fuzzy logic-driven deep learning framework for the automated detection of fake reviews. The proposed approach integrates a Convolutional Neural Network (CNN) for sentiment feature extraction and classification with a Long ShortTerm Memory (LSTM) network for capturing contextual and sequential dependencies within review texts. The outputs generated by both deep learning models are combined using a fuzzy logic inference system, which enhances classification accuracy while improving the interpretability of detection results. The framework is implemented using Python, Flask, TensorFlow, and MySQL, with Natural Language Toolkit (NLTK) employed for text preprocessing and feature preparation. The proposed model was evaluated using the Amazon Food Reviews dataset, where it demonstrated superior performance compared to individual CNN and LSTM models. Experimental results indicate that the hybrid architecture effectively identifies deceptive reviews while maintaining high accuracy, precision, recall, and F1-score values. In addition, a web-based interface was developed to enable retailers and platform administrators to analyze review data, classify reviews, and detect fraudulent content in real time. The findings suggest that the proposed hybrid fuzzy deep learning framework provides an accurate, scalable, and interpretable solution for enhancing trust and reliability in online review systems.

Keywords : Fake Review Detection, Deep Learning, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Fuzzy Logic, Sentiment Analysis, Opinion Mining, E-Commerce Security.

Author : M Venunath, G Santhosh Kumar, V Subhashini

Title : A Multi-Feature Hybrid Fuzzy Logic and Deep Learning Framework for Accurate Fake Review Classification

Volume/Issue : 2026;03(02)

Page No : 124-138