Journal of Science Engineering Technology and Management Science

Zero-day attacks represent one of the most critical challenges in cybersecurity, exploiting previously unknown vulnerabilities for which no signatures exist. Traditional detection mechanisms, including signature-based and anomaly-based systems, fail to identify these novel threats, leaving systems vulnerable for extended periods (averaging 312 days between exploitation and discovery). This paper introduces a groundbreaking hybrid framework for zero-day attack detection that synergistically combines Quantum Machine Learning (QML) with classical rule-based systems, deployed through a comprehensive Python Tkinter application. The core innovation lies in using Geometric Quantum Machine Learning (GQML) to train on known attack patterns and normal network behavior, creating a quantum-enhanced feature space that can identify deviations indicative of zero-day exploits. The system architecture comprises four integrated modules: (1) a real-time network monitoring engine capturing 47 distinct flow metrics at 10ms granularity, (2) a GQML classifier utilizing parameterized quantum circuits with 8 qubits and 64-dimensional feature encoding, (3) a rule-based expert system incorporating 156 SNORT-derived signatures for known attack validation, and (4) a Tkinter-based graphical interface providing real-time visualization, alert management, and interactive analysis. The GQML model achieves 96.8% accuracy in distinguishing known attacks from normal traffic, while the hybrid decision framework demonstrates 89.3% detection rate for zero-day variants with only 3.2% false positives—significantly outperforming classical approaches (SVM: 71.4%, Random Forest: 76.8%, LSTM: 82.1%) on the CIC-IDS2017 and CSE-CIC-IDS2018 datasets augmented with 15 synthetic zero-day scenarios. The Tkinter application provides network administrators with intuitive dashboards showing real-time traffic analytics, quantum confidence scores, rule violation alerts, and historical trend analysis. This work represents the first integration of quantum machine learning with practical cybersecurity tools, demonstrating that quantum advantages can be realized in production environments for detecting previously unseen threats.

Keywords : Zero-Day Attack, Quantum Machine Learning, GQML, Tkinter, Network Monitoring, Hybrid Detection, Quantum Neural Networks, Cybersecurity

Author : Dr. K. Anuradha,Vanamala Vasavi,Vooturi Kamali,Rachapally Chandra Vamshi,Polepongu Vinusha,Sandha Vamshi

Title : Zero-Day Attack Detection Using Quantum Machine Learning: A Hybrid Framework

Volume/Issue : 2026;03(03)

Page No : 439-449

Data leakage poses significant threats to organizational security, with traditional detection methods requiring access to sensitive information in plaintext. This paper presents a novel framework for privacy-preserving data leakage detection that combines fully homomorphic encryption (FHE) with deep learning architectures. Our approach enables organizations to detect sensitive data exposure without decrypting the monitored content, addressing critical privacy concerns. We introduce HE-DLDNet, a modified transformer-based neural network that operates directly on encrypted data streams, achieving 94.3% detection accuracy while maintaining computational efficiency. Experimental results on benchmark datasets demonstrate that our method reduces computational overhead by 73% compared to existing homomorphic approaches while preserving privacy guarantees. The framework supports real-time detection with latency under 250ms, making it suitable for production deployment. This work bridges the gap between strong privacy preservation and practical data leakage detection systems.

Keywords : Homomorphic encryption, deep learning, data leakage detection, privacy preservation, network security

Author : Mrs. P. Revathy,Thodendula Madhusha Yadav,Tambaku Swaroopa Rani,Oduri Sathya Kiran,Pathri Poojitha,Vandana Ravi Teja

Title : Privacy-Preserving Data Leakage Detection Using Homomorphic Encryption and Deep Learning

Volume/Issue : 2026;03(03)

Page No : 450-457

The increasing demand for efficient resource management and fair billing mechanisms in shared living spaces, hostels, rental properties, and co-working environments has necessitated the development of automated pay-for-use systems. Traditional flat-rate billing models often lead to inequitable cost distribution and resource wastage, as users lack incentives for conservation. This paper presents an RFID-based Resource Billing System, a comprehensive IoT-enabled pay-for-use platform that tracks individual consumption of resources (electricity, water, TV, fans, appliances) and generates usage-based bills with integrated data analytics. The system comprises three integrated components: (1) an ESP8266-based hardware module with RFID readers (MFRC522) and relay circuits connected to resource outlets, enabling tap-to-start and tap-to-stop usage tracking with 99.8% accuracy [5]; (2) a Firebase Realtime Database cloud infrastructure that maintains user profiles, resource rates, usage sessions, billing records, and synchronization across multiple devices with 150ms average latency [6]; and (3) a Python Tkinter analytics application that performs comprehensive data analysis including usage pattern recognition, peak demand forecasting, anomaly detection, and customizable report generation (daily, weekly, monthly, yearly). The system implements a complete workflow: user registration with RFID card assignment, resource selection via RFID tap, realtime usage tracking with session management, automatic billing based on configurable rates ($0.12/kWh for electricity, $0.05/gallon for water, $0.50/hour for TV, etc.), and payment processing. The ESP8266 firmware, written in Arduino C++, maintains persistent connection to Firebase using REST APIs, with watchdog timers ensuring 99.9% uptime.The Python analytics application utilizes pandas for data manipulation, matplotlib/seaborn for visualization, scikit-learn for predictive modeling, and tkinter for the graphical user interface. Experimental deployment across 50 rooms in a university hostel over 6 months generated 25,000+ usage sessions, demonstrating 34% reduction in overall resource consumption compared to flatrate billing, 28% cost savings for low-usage users, and 97.5% user satisfaction. The analytics module identified peak usage patterns (evenings 6-10 PM accounting for 45% of electricity consumption), detected anomalous usage events (15 cases of potential theft/misuse), and forecasted demand with 92% accuracy using SARIMA models. The system supports up to 500 concurrent devices per Firebase instance, with daily backups and GDPR-compliant data retention policies. This work represents the first integrated RFID-based pay-for-use system combining real-time IoT tracking with advanced analytics, demonstrating significant potential for equitable resource billing and consumption optimization

Keywords : RFID, IoT, Pay-for-Use, ESP8266, Firebase, Resource Billing, Data Analytics, Python, Tkinter, Usage Tracking, Smart Metering, Consumption Analysis

Author : Dr. N. Srinivasa Rao,Swamulapally Ravindra,Konda Archana,Valleti Venkata Lakshmi,Veeramsetti Manikanta,Chetla Saikiran

Title : RFID-Based Resource Billing: A Pay-for-Use IoT System

Volume/Issue : 2026;03(03)

Page No : 458-467

Air pollution has emerged as one of the most critical environmental challenges of the 21st century, with the World Health Organization estimating that 99% of the global population breathes air exceeding safe quality limits, resulting in approximately 7 million premature deaths annually. Industrial emissions, vehicle exhaust, and biomass burning release particulate matter (PM2.5 and PM10) and toxic gases that penetrate deep into the respiratory system, causing cardiovascular and respiratory diseases. This paper presents a novel autonomous robotic vehicle designed to detect, capture, and remove smoke pollutants from the air using electrostatic precipitation technology combined with IoT-based control systems. The vehicle integrates an ESP8266 microcontroller for autonomous navigation and control, an MQ2 gas sensor for real-time smoke detection, and a custom-designed electrostatic precipitator (ESP) powered by a 400kV pulse generator for highefficiency particulate removal. The robotic platform utilizes a motor driver IC with four DC motors, enabling omnidirectional movement toward pollution sources when detected. Upon smoke detection above threshold levels (500 ppm), the vehicle halts navigation and activates the electrostatic precipitation system, which generates a high-voltage corona discharge (400kV) between an aluminum tube chamber and a central electrode, ionizing airborne particles and collecting them on oppositely charged plates. The electrostatic precipitator achieves 94.7% removal efficiency for PM2.5 particles and 96.2% for PM10 particles at an airflow rate of 0.5 m³/s, as validated through controlled chamber testing. The pulse generator circuit, based on a flyback transformer topology with MOSFET switching at 25kHz, produces the required 400kV potential while consuming only 45W of power, making it suitable for battery-powered mobile applications. The ESP8266 microcontroller manages sensor data acquisition, motor control logic, and high-voltage system activation through optoisolated relay circuits for safety isolation. Experimental evaluation across 50 test runs in controlled smoke environments demonstrates an average smoke detection response time of 2.3 seconds, with the vehicle successfully navigating to pollution sources and reducing ambient particulate concentration by 78% within 5 minutes of operation. The system achieves 94.2% accuracy in distinguishing smoke from other airborne particles using differential sensor fusion techniques. Power consumption analysis shows the vehicle can operate continuously for 2.5 hours on a 12V, 20Ah battery bank, sufficient for typical urban pollution hotspot patrolling. The total material cost of $185 per unit makes this technology accessible for deployment in industrial zones, urban hotspots, and developing regions where stationary air purifiers are impractical. This work demonstrates that mobile electrostatic precipitation combined with autonomous navigation offers a viable, scalable solution for targeted air pollution control, contributing to environmental health and sustainable urban development.

Keywords :

Author : Mr. Dandu Srinivas,Som Rithika,Kolguri Shirisha,Roopani Aparna,Vennam Akshara

Title : Design and Development of an Autonomous Smoke Pollution Removing Vehicle Using Electrostatic Precipitation

Volume/Issue : 2026;03(03)

Page No : 468-474

Heavy construction infrastructure including bridges, dams, high-rise buildings, and industrial facilities requires regular structural health monitoring to ensure safety and prevent catastrophic failures. Traditional manual inspection methods are labor-intensive, time-consuming, and often require scaffolding or specialized access equipment that puts inspectors at risk. This paper presents a novel drone-based crack analysis system implemented as a Python Windows application that processes real-time video feeds and sensor data from unmanned aerial vehicles (UAVs) to automatically detect, classify, and analyze cracks in heavy construction structures. The system integrates a custom-trained YOLOv8 deep learning model using a comprehensive dataset of 25,000 annotated images from Roboflow, encompassing diverse crack types including hairline cracks, structural cracks, surface cracks, fatigue cracks, and thermal cracks across various construction materials (concrete, steel, masonry). The architecture comprises four integrated modules: (1) Drone Communication Module that interfaces with DJI and custom UAV platforms via MAVLink protocol, capturing synchronized 4K video at 30fps and telemetry data including GPS coordinates, altitude, orientation, and timestamps; (2) Real-Time Video Processing Pipeline implementing frame extraction, enhancement through adaptive histogram equalization and noise reduction filters, and parallel processing using GPU acceleration; (3) YOLO-Based Crack Detection Engine featuring YOLOv8m architecture with 3.2 million parameters, achieving mean Average Precision (mAP) of 0.943 at 0.5 IoU threshold and inference speed of 45 frames per second on NVIDIA RTX 3060 GPU; (4) Crack Analysis and Reporting Module that quantifies crack dimensions (length, width, depth estimation), classifies severity into four levels (minor, moderate, severe, critical), analyzes potential causes based on crack patterns (structural stress, thermal expansion, material fatigue, settlement), and generates comprehensive inspection reports with geotagged crack locations. The model was trained on Roboflow datasets comprising 25,000 images with 45,000 annotated crack instances, augmented with synthetic variations (rotation, scaling, brightness adjustment, Gaussian noise) to improve generalization. Experimental evaluation on 50 real-world construction sites demonstrates detection accuracy of 96.8%, precision of 95.2%, recall of 94.7%, and F1-score of 0.949. The system successfully identifies cracks as small as 0.5mm width from drone altitudes up to 15 meters. Sensor data fusion incorporating infrared thermal imaging detects subsurface cracks invisible to optical cameras, improving early detection by 34%. The Windows application provides an intuitive GUI with real-time visualization, historical data comparison, automated report generation in PDF/Excel formats, and cloud synchronization for collaborative analysis. This work represents the first integrated drone-based crack analysis system combining state-of-the-art YOLO deep learning with multi-sensor data fusion for comprehensive heavy construction structural health monitoring.

Keywords : Drone-based Inspection, Crack Detection, YOLOv8, Deep Learning, Heavy Construction, Structural Health Monitoring, UAV, Computer Vision, Roboflow Dataset, Python Windows Application

Author : Mr. Mohd Nawazuddin,Gollapally SriVarshitha,Begari Rugvedhini,Koneru Sravan Kumar,Routh Rahul

Title : Drone-Based Heavy Construction Crack Analysis Using Deep Learning: A YOLO-Based Framework

Volume/Issue : 2026;03(03)

Page No : 475-484

In the digital era, the widespread use of mobile communication has made Short Message Service (SMS) a prime target for spammers and cybercriminals. Spam messages not only disrupt user experience but often serve as vectors for phishing attacks, malware distribution, and fraudulent schemes. With the proliferation of such threats, there is a pressing need for intelligent systems capable of automatically detecting and filtering spam content to safeguard users from potential harm.This project presents a hybrid machine learning approach that addresses two critical tasks: SMS spam detection and URL malicious classification. The first component focuses on classifying SMS messages as either spam or ham (legitimate) using natural language processing (NLP) techniques and supervised machine learning algorithms. Text preprocessing methods such as tokenization, stopword removal, and TF-IDF vectorization are employed to transform raw SMS text into meaningful features suitable for model training.The second component targets the classification of URLs embedded within SMS messages to determine whether they are malicious or benign. By extracting lexical features—such as URL length, number of digits, use of special characters, and domain-related attributes—the system utilizes ensemble classifiers like Random Forest and XGBoost to detect suspicious URLs. This duallayered detection mechanism enhances security by identifying both unsolicited messages and hidden threats within them

Keywords :

Author : Mrs. J.Raja Kala,Palem Hari Keerthi,Nayudu Deepika, Sambaru Venkata Sai Kalyan,Jupalli Sriyagna

Title : Intelligent Sms Spam Detection And Malicious Url Classification Using Machine Learning

Volume/Issue : 2026;03(03)

Page No : 485-491

TheIn the rapidly evolving landscape of cybersecurity, timely and efficient sharing of threat intelligence is crucial for organizations to defend against sophisticated cyber-attacks. This project focuses on the implementation of a Threat Intelligence Sharing Platform using the Malware Information Sharing Platform (MISP), an open-source tool designed to facilitate the collection, sharing, and analysis of cyber threat data. By leveraging MISP, organizations can collaboratively share indicators of compromise (IOCs), attack patterns, and threat actor profiles to improve overall situational awareness and response capabilities. To enhance the effectiveness of the platform, this project integrates machine learning techniques to automate the detection, classification, and prioritization of threat data shared within the MISP ecosystem. Machine learning models analyze vast amounts of threat intelligence to identify patterns and anomalies that might indicate emerging threats or false positives, thereby improving the accuracy and speed of threat detection. This approach aims to reduce the manual effort required for threat analysis and enables proactive defense strategies. The implementation includes developing data ingestion pipelines to normalize and preprocess diverse threat intelligence formats, ensuring seamless integration with machine learning modules. Various supervised and unsupervised learning algorithms are explored for tasks such as threat classification, clustering similar incidents, and predicting attack trends. The platform also supports continuous learning, allowing models to evolve as new threat data becomes available, thereby maintaining high detection efficacy over time. Evaluation of the platform involves benchmarking the performance of machine learning models on real-world threat intelligence datasets, measuring metrics such as detection accuracy, false positive rates, and processing latency. The results demonstrate significant improvements in identifying and correlating threats compared to traditional rule-based methods. Additionally, the platform promotes collaborative cybersecurity by enabling organizations to share actionable insights securely and efficiently. In conclusion, this project presents a comprehensive solution for enhancing threat intelligence sharing through the integration of MISP and advanced machine learning techniques. By automating threat analysis and improving data sharing, the platform empowers cybersecurity teams to respond faster and more effectively to emerging cyber threats, ultimately strengthening the security posture of participating organizations.

Keywords :

Author : Mr. Ch. Siva Prakash, Mallelli Sai Manasa,Bejjanki Sukanya Sowmya, Ganthala Ajaykumar, Eeda Uday Kiran

Title : Design and Development of a Cyber Threat Intelligence Sharing Platform Using MISP

Volume/Issue : 2026;03(03)

Page No : 403-497

In the evolving landscape of digital communication and data storage, cybersecurity remains a paramount concern. This project proposes a hybrid security framework that combines Hybrid Encryption (AES + ECC) with Image-based Steganography to provide enhanced protection for user data stored on centralized or decentralized servers. The hybrid encryption approach ensures that data cannot be decrypted even if malicious actors obtain partial keys, as it uses both symmetric and asymmetric algorithms. Additionally, Steganography conceals sensitive messages within image files, enabling covert data transmission while maintaining the appearance of innocuous media. To further ensure data integrity, the system generates a unique hashcode for each uploaded file, allowing verification at any time.Access control is fortified through multi-factor authentication, combining traditional credentials with OTPbased email verification. Beyond security operations, the platform also includes user education tools, providing learning materials and real-time cybersecurity news updates. Developed using Python and MySQL, the application empowers users to encrypt files, hide data in images, retrieve decrypted files, and stay informed about modern threats—all through a secure and interactive web interface

Keywords : Hybrid Encryption, AES, ECC, Steganography, Data Security, Cybersecurity, Hashcode Verification, Multi-Factor Authentication, OTP Verification, Secure Data Storage, Image-based Data Hiding, Python, MySQL.

Author : Mr. M. Mahesh, Gopi Dhanalakshmi, Punem Swarupa, Daggupati Ganesh, Pakanati Jaswanth

Title : Secure Data Protection Using Hybrid Encryption and Steganography Techniques

Volume/Issue : 2026;03(03)

Page No : 498-504

In recent years, lifestyle-related health disorders such as cardiovascular diseases, stress, obesity, and hypertension have increased significantly due to irregular sleep habits and abnormal physiological conditions. This project proposes a predictive system that analyzes sleep patterns and pulse rate data to assess potential health risks. By collecting real-time or historical data from wearable devices or health monitoring systems, the model identifies patterns associated with unhealthy lifestyles. Machine learning techniques are employed to classify individuals into different risk categories based on features such as sleep duration, sleep quality, resting pulse rate, and variability. The system aims to provide early warnings and personalized insights, enabling users to take preventive measures and adopt healthier routines. This approach not only enhances awareness but also contributes to proactive healthcare management, reducing the likelihood of severe health complications

Keywords : Lakkadasu Sai Laxmi Ganesh, 3 Shaik Subhani, 4Kondapalli Harshitha, 5 P R Venkata Sai Phani Kumar 1Assistant Professor, Department of Computer Science & Engineering, Sai Spurthi Institute Of Technology 2,3,4,5B. Tech Students, Department of Computer Science & Engineering, Sai Spurthi Institute Of Technology ABSTRACT In recent years, lifestyle-related health disorders such as cardiovascular diseases, stress, obesity, and hypertension have increased significantly due to irregular sleep habits and abnormal physiological conditions. This project proposes a predictive system that analyzes sleep patterns and pulse rate data to assess potential health risks. By collecting real-time or historical data from wearable devices or health monitoring systems, the model identifies patterns associated with unhealthy lifestyles. Machine learning techniques are employed to classify individuals into different risk categories based on features such as sleep duration, sleep quality, resting pulse rate, and variability. The system aims to provide early warnings and personalized insights, enabling users to take preventiv

Author : Mrs. J.Raja Kala, Lakkadasu Sai Laxmi Ganesh, Shaik Subhani, Kondapalli Harshitha, P R Venkata Sai Phani Kumar

Title : PREDICTING LIFE STYLE RELATED HEALTH RISK USING SLEEP PATTERN AND PULSE

Volume/Issue : 2026;03(03)

Page No : 505-510

The rapid spread of fake news on digital platforms has become a serious challenge, influencing public opinion and decision-making. This project presents an Intelligent Fake News Detection Framework using an Optimized Multi-Support Vector Machine (MultiSVM) classification approach to accurately identify fake and genuine news. The proposed system performs text preprocessing, feature extraction, and classification using multiple SVM models combined for improved prediction accuracy. Optimization techniques are applied to select the best parameters and features, reducing misclassification and improving performance. Experimental results show that the optimized Multi-SVM framework achieves higher accuracy, precision, and reliability compared to traditional single-classifier models. This framework can be effectively used in real-world applications such as social media monitoring, online news verification, and misinformation control.

Keywords : Fake News Detection, Machine Learning, Multi-SVM Classification, Text Mining, Feature Extraction, Optimization Techniques, Natural Language Processing (NLP)

Author : Mr. Ch. Bala Krishna, D V Sai Gayatri Sriprada, Thota Deepika, Khan Shabana Thabassum, Sathenapalli Phanindra

Title : An Intelligent Fake News Detection Framework Using Optimized Multi-SVM Classification

Volume/Issue : 2026;03(03)

Page No : 511-516

The restoration of marker-occluded hematoxylin and eosin (H&E) stained whole slide histology images is a critical task in digital pathology, as annotations and markings made by pathologists often obscure important tissue structures. This work proposes a deep learning-based approach utilizing Generative Adversarial Networks (GANs) to automatically reconstruct and restore occluded regions in histopathological images. The model is trained on paired datasets of clean and artificially occluded H&E images to learn the underlying tissue distribution and generate realistic reconstructions. The generator network predicts the missing regions, while the discriminator ensures the authenticity and structural consistency of the restored output. The proposed method effectively removes artifacts such as pen marks and labels while preserving cellular morphology and diagnostic features. Experimental results demonstrate improved visual quality and structural similarity compared to traditional inpainting techniques, making it a promising solution for enhancing the reliability of digital pathology workflows.

Keywords : Generative Adversarial Networks (GANs), Histopathology Image Restoration, Hematoxylin and Eosin (H&E) Staining, Whole Slide Imaging (WSI), Image Inpainting, Marker Occlusion Removal, Digital Pathology, Deep Learning, Medical Image Processing, Tissue Structure Preservation, Image Reconstruction, Computer-Aided Diagnosis (CAD),Biomedical Image Analysis, Artifact Removal.

Author : Mr Sarma Adithe, Ch.Sharmila, K. Kusuma, S. Yamuna Devi, G. Roja

Title : RESTORATION OF MARKER OCCLUDED HEMATOXYLIN AND EOSIN-STAINED WHOLE SLIDE HISTOLOGY IMAGES USING GENERATIVE ADVERSARIAL NETWORKS

Volume/Issue : 2026;03(03)

Page No : 524-530

The filter design optimization (FDO) problem is defined as finding a set of filter coefficients that yields a filter design with minimum complexity, satisfying the filter constraints. It has received a tremendous interest due to the widespread application of filters. Assuming that the coefficient multiplications in the filter design are realized under a shift-adds architecture, the complexity is generally defined in terms of the total number of adders and subtractors. In this paper, we present an exact FDO algorithm that can guarantee the minimum design complexity under the minimum quantization value, but can only be applied to filters with a small number of coefficients. We also introduce an approximate algorithm that can handle filters with a large number of coefficients using less computational resources than the exact FDO algorithm and find better solutions than existing FDO heuristics. We describe how these algorithms can be modified to handle a delay constraint in the shift-adds designs of the multiplier blocks and to target different filter constraints and filter forms. Experimental results show the effectiveness of the proposed algorithms with respect to prominent FDO algorithms and explore the impact of design parameters, such as the filter length, quantization value, and filter form, on the complexity and performance of filter designs. Further, this project is enhanced by using Look Up Table approach. APC and OMS binary properties are used for implementing LUT based FIR filter.

Keywords : Anti Symmetric Product Coding (APC), Odd multiple Storage(OMS), Look Up Table (LUT), Finite Impulse Response(FIR).

Author : Dr.Subransu Padhee, S. Sathvika, G. V. V. Pavani, N. Bhanu Sri, Ch. Naga Sai

Title : IMPLEMENTATION OF MULTIPLIER ARCHITECTURE USING EFFICIENT CARRY SELECT ADDERS FOR SYNTHESIZING FIR FILTERS

Volume/Issue : 2026;03(03)

Page No : 531-537

Consumption of Energy is the major factor, in the various processing application like DSP, ASIC, and FPGA. This project presents power optimized more accurate and approximate 8 x 8 multiplication architecture design. To achieve high accura,4cy, this concept utilizes accurate (i.e., exact) 4:2 compressors in the higher significance weights. To reduce power consumption, high-order approximate compressors are used in the middle significance weights. As an enhancement of this concept, latency and density optimized modified round square architecture is proposed with more accuracy. The approach is to round the operands to the nearest exponent of two. This way the computational intensive part of the multiplication is omitted improving speed and energy consumption at the price of a small error. The proposed approach is applicable to both signed and unsigned multiplication.

Keywords : Digital signal processing, Arithmetic Logic Unit, multiplier and accumulate unit, elliptic curve cryptography, Approximate Multiplier

Author : Dr. SIVA SANKARA PHANI TAMMIREDDY,Gudimetla Harika, Pilla Sairam,Chandana Hari, Kopella Venkata Naga Abhinayak

Title : LOW POWER HIGH ACCURACY APPROXIMATE MULTIPLER USING HIGH ORDER COMPRESSOR

Volume/Issue : 2026;03(03)

Page No : 538-548

This paper proposes a 4-bit array multiplier useful in the design of basically mixer circuit, which is highly involved in signal and image processing using an efficient low-power VLSI technique. The presented architecture is completely implemented adiabatic techniques in the Near Threshold Region, which optimize the product of propagation delay and power dissipation. Multiplier is the most frequently used element in many digital electronics applications. Depending on the applications, various types of multipliers emerge. With this technique, the total power dissipation, i.e. dynamic power dissipation as well as static power dissipation is less as compared to the conventional CMOS technique. The Near Threshold Adiabatic Logic (NTAL) technique is used with a single time varying power supply which reduces the clock tree management and enhances the energy-saving capability. Simulation of the proposed design is done by Tanner tools schematic editor with wedit simulator on TSMC 65nm technology node to verify the optimized result.

Keywords : Near Threshold Adiabatic Logic, Adiabatic, CMOS Multiplier, efficiency, Power leakage.

Author : Venkata Satish Dhulipudi, Sundaraneedi Sai Krishna Kumari, Nandra Vijaya Lakshmi, Kallakuri Kavya Sri, veera Hyma NAGALATHA

Title : DESIGN AND OPTIMIZATION OF 4-BIT ARRAY MULTIPLIER WITH ADIABATIC LOGIC USING 65?NM CMOS TECHNOLOGIES

Volume/Issue : 2026;03(03)

Page No : 549-558

The increasing concerns over the safety of women and children have led to the need for smart, realtime security solutions. This project proposes a Smart Intelligent Security System using Raspberry Pi that integrates IoT, GPS, and communication technologies to provide immediate assistance during emergencies. The system is designed to detect distress situations through manual triggers (panic button) or automatic sensing mechanisms. Upon activation, the device captures location data using GPS and sends alert messages to predefined contacts via GSM or internet-based services. Additionally, features such as live audio/video streaming, real-time tracking, and emergency siren activation enhance the system’s effectiveness. The compact and cost-effective design makes it suitable for wearable or portable safety devices. This solution aims to ensure quick response, continuous monitoring, and improved personal safety.

Keywords : Women Safety System, Child Security, Raspberry Pi, Internet of Things (IoT), GPS Tracking, GSM Communication, Panic Button, Real-Time Monitoring, Embedded Systems, Emergency Alert System.

Author : 1VAYARA PRASANNA LAKSHMI, ANKANI AAKANKSHA, LAKKIMSETTY SATYA NAGA VENKATA RAMA SOWMYA, AMBATI BHARGAVI, PATI VENKATA SATYA SIVA

Title : SMART INTELLIGENT SYSTEM FOR WOMEN AND CHILD SECURITY

Volume/Issue : 2026;03(03)

Page No : 559-564

The proposed system is an IoT-based Patient Health Monitoring System designed to continuously track essential physiological parameters of a patient in real time. It incorporates a heartbeat sensor, blood pressure sensor, and ECG sensor to monitor pulse rate, blood pressure, and cardiac signals, while a DHT11 sensor measures the surrounding temperature and humidity. All sensor data is collected and processed by an Arduino UNO, which acts as the main control unit of the system. A NodeMCU (ESP8266) module transmits the processed data to the ThingSpeak cloud platform, enabling doctors and caregivers to remotely access and monitor the patient’s health condition from any location. For local observation, a 16×2 LCD is used to display live readings at the patient’s side. In addition, a buzzer is integrated to provide immediate alerts whenever any of the measured parameters exceed preset threshold values, ensuring quick response during emergencies. The system operates on a regulated 5V power supply supported by a 12V 1A adapter to guarantee stable performance. This smart solution minimizes the need for continuous manual checking, enhances patient safety, and is highly suitable for hospitals, home healthcare, and elderly care monitoring.

Keywords : IoT-Based Health Monitoring, Arduino UNO, Remote Patient Care, Physiological Sensors, Cloud Data Transmission

Author : Mr.B.Vamsy Krishna,S. Abhinaya Sri,P. Jeevana Sree,S. Poorneswari,Ch. Kalyan Babu

Title : Design of patient health monitoring system based on iot

Volume/Issue : 2026;03(03)

Page No : 565-573

This paper presents the design of a high-speed 16-bit synchronous Gray code counter with a constant counting rate, aimed at overcoming the performance degradation observed in conventional binary counters as bit-width increases. The proposed architecture employs a pre-scaled counter structure that maintains a uniform counting rate independent of counter size. An early state detection mechanism reduces critical path delay by enabling faster generation of control signals for subsequent sub-counters. Johnson counters replace traditional T flipflops, reducing propagation delay and improving operating speed. A flipflop-based clock gating technique selectively activates the clock only when required, thereby reducing unnecessary switching activity and dynamic power consumption without compromising speed. Results on Xilinx Artix-7 FPGA demonstrate 91.8% total power reduction (0.131 W vs. 1.6 W) and near-complete elimination of dynamic power (<0.001 W), with improved junction temperature (25.2°C vs. 28.0°C), making the design suitable for high-speed digital applications including frequency dividers, phase-locked loops, and IoT-based systems.

Keywords : Gray code counter; constant time counter; high-speed counter; pre-scaled counter; flip-flop based clock gating; FPGA; Xilinx Artix-7; dynamic power reduction; Johnson counter; synchronous design

Author : Mrs. V. Priyadarshini, Nethala. Durga Devi, Shaik. Nasreen Chandini,Veenam. Geetha kumari, Vinukonda. Vikram

Title : A Flip-Flop Clock Gated 16-Bit Synchronous Gray Code Counter Architecture for Constant Rate High-Speed Counting

Volume/Issue : 2026;03(03)

Page No : 574-577

The construction industry has developed at high rates and the resources consumed rapidly are natural river sand and cement, which cause a severe degradation of nature and also lead to high carbon dioxide emissions. At the same time, the amounts of industrial waste products including granite powder, waste glass, fly ash, and marble powder produced by stone-processing and manufacturing industry pose severe disposal and health problems. The research examines the potential of using granite powder as a partial substitute of natural fine aggregate and glass powder as an auxiliary cementitious material to come up with sustainable and strong concrete. Experimental studies were conducted by substituting cement with up to 25% of natural sand with glass powder and then substituting natural sand with granite powder up to 50% and a combination of the two. Moreover, fly ash (20 percent) was added to enhance optimum mixes and marble powder (10 percent) was incorporated as a filler material. Slump tests were used to evaluate fresh property and the hardened property that were assessed included compressive strength, flexural strength, splitting tensile strength, density, abrasion resistance and modulus of elasticity. The properties of durability such as water absorption, water penetration, acid attack, sulphate attack and quick penetration through chloride were also evaluated. EMU Microstructural characterization was done by SEM and XRD to learn the behavior of the internal matrix and hydration. Findings show that the ideal ratios of granite powder and the glass powder are influential in improving the mechanical performance and durability as a result of better particle packing and pozzolanic reactions as well as reducing the reliance on traditional materials. The research shows that mixed use of these waste products is an effective, environmentally friendly and sustainable option towards concrete production

Keywords : Granite powder, waste glass, fly ash, marble powder, concrete.

Author : Menda Sai Srinivas Varma, M.Suresh Chandra Babu

Title : Performance Evaluation of Eco-Friendly Concrete Using Granite Powder And Waste Glass As Partial Replacements

Volume/Issue : 2026;03(03)

Page No : 578-587

This paper is an experimental study on the effect of steel fibers as such, polypropylene fibers as such and when combined in hybrids on fresh and hardened properties of M40 grade concrete. The main aim is to test the workability, compressive strength, split tensile strength and flexural behaviour of concrete to the different types of fibres and volume fractions of the fibres. An equal mixture of crimped steel fibers and polypropylene fibers were used together and separately to prepare concrete mixtures at different proportions. The fresh concrete properties were also evaluated using slump and compaction factor tests, the results of which showed that workability decreased with fiber content especially the steel fiber-dominant mixes. The hardened concrete properties were tested at the following curing times namely, 7 days, 14 days and 28 days. The experimental evidence showed that there were significant differences in compressive and split tensile strengths of all fiber-reinforced mixes in comparison to conventional concrete with hybrid fiber-reinforced concrete showing the highest increase in strength. Flexural tests on beam specimens have shown improved resistance to cracks, increased ultimate load capacity and improved deflection control particularly in fiber hybrids. The interaction between steel and polypropylene fibers was found to enhance the distribution of stresses, the introduction of cracks, and ductility. On the whole, the results prove that a hybrid type of fiber reinforcement has a substantial positive effect on the mechanical performance and structural behavior of concrete and can be used in the context of better durability, strength, and crack resistance.

Keywords : Hybrid Fiber Reinforced Concrete, Steel Fibers, Polypropylene Fibers, Compressive Strength, Flexural Performance, Workability.

Author : Sureddi sindhura, C. Venkata Siva Rama Prasad

Title : Performance Evaluation of Steel–Polypropylene Hybrid Fiber Reinforced Concrete

Volume/Issue : 2026;03(03)

Page No : 588-596

The rapid growth of blockchain technology and cryptocurrencies has created new opportunities for digital financial transactions. However, the anonymity and decentralized nature of blockchain networks also make them attractive platforms for illegal financial activities such as money laundering. Detecting suspicious transactions within blockchain systems has become a major challenge for regulatory authorities and financial institutions. This study proposes a secure blockchain transaction analysis framework using machine learning techniques to detect potential money laundering activities. The system analyzes blockchain transaction data and extracts important features such as transaction frequency, wallet interactions, transaction volume, and network patterns. These features are then processed using machine learning algorithms to classify transactions as legitimate or suspicious. The proposed framework integrates data preprocessing, feature extraction, and supervised machine learning models such as Random Forest, Support Vector Machine (SVM), and Decision Tree to improve detection accuracy. By analyzing large-scale blockchain transaction datasets, the system can identify abnormal patterns and detect laundering activities more effectively than traditional rule-based methods. The results demonstrate that machine learning-based analysis can significantly enhance the efficiency, accuracy, and scalability of anti-money laundering systems in blockchain environments. The proposed approach provides a secure and intelligent solution for monitoring cryptocurrency transactions, helping financial institutions and regulatory agencies prevent illegal financial activities and maintain transparency in digital financial ecosystems.

Keywords :

Author : RAVULA JEEVITHA, A JYTHOSNA, BABARIYA JAGADISH KUMAR, BALAGA MANOHAR BABU, BANDI SNEHA

Title : SECURE BLOCKCHAIN TRANSACTION ANALYSIS FOR MONEY LAUNDERING DETECTION USING MACHINE LEARNING

Volume/Issue : 2026;03(03)

Page No : 597-602

Nurses experience high levels of occupational stress due to long shifts, critical patient care, and emotional demands. In India, studies indicate that nearly 60% of healthcare professionals report moderate to high stress levels, contributing to burnout and reduced efficiency. The objective of this research is to predict nurse stress in real-time using wearable sensor data and iterative learning methods. It aims to provide accurate stress classification for timely health interventions. Traditionally, nurse stress monitoring relies on self-reported questionnaires, periodic interviews, and manual observation of behavioral cues. Supervisors assess workload, fatigue, and stress levels subjectively, often resulting in delayed interventions and incomplete data. Manual stress assessment is prone to bias, inaccuracies, and inconsistencies. It cannot capture real-time fluctuations, often misses early warning signs, and is resourceintensive. Consequently, interventions are reactive rather than preventive. The motivation is to overcome limitations of manual assessment by using wearable sensors and machine learning to provide continuous, objective, and accurate stress monitoring. This ensures timely interventions, reduces burnout, and improves nurse productivity and patient care outcomes. The proposed system integrates wearable physiological sensors with machine learning models. Logistic Regression provides a baseline probabilistic stress prediction, while Ridge Classifier handles noisy signals with regularization. The MLP–ETC hybrid combines deep feature extraction with ensemble-based classification for higher accuracy. Iterative learning continuously updates models with new data, enabling real-time stress monitoring and health interventions. Machine learning helps identify stress patterns early, predict risk levels, and support personalized recommendations, improving nurses’ mental health and operational efficiency

Keywords :

Author : PAINDLA THIRUPATHI, ARRABOINA AKHIL, BEEREDDY SRUJANA, DUVVAKA VAMSHI KRISHNA, B MANASA

Title : SMART HEALTHCARE SYSTEM FOR NURSE STRESS PREDICTION USING WEARABLE SENSORS AND MACHINE LEARNING

Volume/Issue : 2026;03(03)

Page No : 603-610

Monitoring hazardous gases in rural and remote areas is essential for protecting human health, agricultural productivity, and environmental safety. However, traditional gas monitoring systems are often limited by high costs, lack of continuous monitoring, and delayed detection capabilities, especially in geographically isolated regions. To address these challenges, this study proposes a predictive monitoring system for hazardous gases using machine learning models that enables real-time detection and early prediction of harmful gas concentrations. The proposed platform integrates gas sensors, data acquisition units, and machine learning algorithms to continuously collect and analyze environmental data such as gas concentration levels, temperature, humidity, and atmospheric conditions. The collected sensor data is processed and used to train machine learning models capable of identifying patterns and predicting potential hazardous gas events before they reach critical levels. Algorithms such as Random Forest, Support Vector Machine, or Artificial Neural Networks can be employed to improve prediction accuracy and system reliability. The system also provides realtime alerts and visualization through a user-friendly monitoring interface, enabling timely preventive actions by local communities, environmental agencies, and agricultural workers. By combining sensor-based monitoring with intelligent predictive analytics, the proposed approach enhances early warning capabilities and supports safer living conditions in rural and remote environments. Overall, this machine learning–based predictive monitoring platform offers a cost-effective, scalable, and intelligent solution for continuous hazardous gas surveillance, contributing to improved environmental monitoring, public safety, and sustainable rural development.

Keywords :

Author : E. GANGARAM, BODAPATI VIGNESH, BUDIDHA HARSHITHA, BHUKYA DINESH, CHADMALA VINOD

Title : PREDICTIVE MONITORING OF HAZARDOUS GASES IN RURAL AND REMOTE AREAS USING MACHINE LEARNING MODELS

Volume/Issue : 2026;03(03)

Page No : 611-616

Real-Time Facial Stress Detection Using Deep Learning is an advanced system designed to analyze facial expressions and detect stress levels and health-related signals in real time. In today’s fast-paced world, mental health issues such as stress, fatigue, and anxiety are increasing, making early detection and monitoring highly important. Traditional methods rely on selfreporting or medical consultation, which may not always be timely or accurate. To address these challenges, this project presents an intelligent facial analysis system using artificial intelligence and deep learning techniques. The proposed system uses computer vision and deep learning models to capture and analyze facial features through a webcam. Technologies such as Open CV and facial landmark detection are used to extract key facial regions, while machine learning models evaluate stress levels, emotional states, and fatigue indicators. The system is capable of identifying parameters such as eye strain, attention levels, facial tension, and mood patterns with good accuracy. To enhance interpretability, the system highlights specific facial regions contributing to stress and provides detailed insights. It also generates personalized suggestions such as breathing exercises, relaxation techniques, and wellness recommendations based on the analysis. The backend is implemented using Python, while the frontend interface provides a userfriendly experience for real-time interaction and visualization of results. By combining artificial intelligence, facial analysis, and real-time processing, Real-Time Facial Stress Detection Using Deep Learning offers a non-invasive, efficient, and accessible solution for monitoring mental well-being. This project demonstrates the potential of AI in healthcare applications, especially in early stress detection, preventive care, and improving overall mental health awareness

Keywords :

Author : SYEDA FATIMA, K. SINDHU, A. JYOTHI, B. LALITHA

Title : Real-Time Facial Stress Detection Using Deep Learning

Volume/Issue : 2026;03(03)

Page No : 691-698

Anaemia is a prevalent hematological disorder requiring early and accurate detection for effective clinical management. Traditional diagnostic methods lack predictive strength, while conventional ML models operate as black boxes limiting healthcare adoption. This paper presents a web-based anaemia prediction system using Django with multiple ML models trained on hematological parameters (RBC, PCV, MCV, MCH, MCHC, RDW, TLC, Platelets, Hemoglobin, Age, Sex). SHAP and LIME explainability techniques are integrated to provide global feature importance and patient-specific local explanations. Among trained models, Random Forest achieves the best performance with 94.2% accuracy. SHAP analysis identifies Hemoglobin, MCV, and MCH as the most influential features. The system bridges the gap between ML accuracy and clinical trust by providing transparent, interpretable predictions.

Keywords : Anaemia Prediction, Explainable AI, SHAP, LIME, Random Forest, Healthcare ML, Django

Author : M. Mohini, A. Akhila, G. Saritha, K. Naveen

Title : Developing a Transparent Anaemia Prediction Model Empowered with Explainable Artificial Intelligence

Volume/Issue : 2026;03(03)

Page No : 617-621

Runtime nondeterminism in software systems causes intermittent failures that are difficult to reproduce and debug. Traditional statistical debugging detects correlations but cannot identify true root causes. This paper implements a Causality-Guided Adaptive Interventional Debugging (AID) system using Django to automatically identify root causes of intermittent software failures. The system analyzes execution logs containing runtime predicates from successful and failed executions, constructs an Approximate Causal Directed Acyclic Graph (AC-DAG), and applies group intervention strategy to iteratively prune non-causal predicates. Experimental evaluation on synthetic and real-world bug datasets shows the system identifies root causes with 92% accuracy while reducing debugging iterations by 65% compared to traditional statistical debugging approaches

Keywords :

Author : G. Pavan, S. Bharathi, K. Srikanya, P. Dileep

Title : Causality-Guided Adaptive Interventional Debugging (AID)

Volume/Issue : 2026;03(03)

Page No : 622-626