Authors - Aakanksha Jain, Tejaskumar Bhatt, Darshita kalyani, Jatin Modh, Abhishek Jain Abstract - The effectiveness of the deep learning models ResNet50, MobileNetV2, VGG16, InceptionV3, and EfficientNetB0 in identifying the freshness of food is evaluated in this study using visual analysis. In resource-constrained situations, these designs are ideal for automated food quality inspection and real-time freshness monitoring as they offer higher accuracy or computational efficiency. Transform learning are used to develop binary classifiers, which were then trained on a dataset of annotated food photos and assessed for efficiency and accuracy. After undergoing standardized preprocessing, the models' capacity to differentiate between fresh and stale food in a variety of test photos was evaluated. The results show the advantages and disadvantages of each model. The current stream of research frequently concentrates on generic picture classification tasks instead of the particular difficulties of distinguishing subtle visual differences in food freshness, leaving a gap in knowledge of model robustness under real-world conditions. The study advances deep learning applications in food quality evaluation by offering useful insights for model selection based on operational requirements.
Tuesday August 25, 2026 9:30am - 11:30am IST Virtual Room AGOA, India
Authors - Jyotiprakash Mishra, Sanjay K. Sahay, Aman Pathak Abstract - RISC-V is an open-source Instruction Set Architecture (ISA) designed with a modular and extensible structure, allowing for customizable implementations. Its simplified base ISA, combined with optional standard and custom extensions, provides flexibility for a wide range of computing applications, from embedded systems to high-performance computing. Its open design accelerates innovation and customization but also introduces security challenges by exposing the architecture to potential attacks. While RISC-V offers significant advantages, its lack of standardized security features compared to proprietary ISAs like ARM and x86 highlights persistent risks, particularly in security-critical applications. For this reason, scrutiny of RISC-V’s security is crucial due to its widespread use in academia and its adoption by countries like India and China, who are looking to benefit from its open nature. We review the current security challenges in RISC-V, examining key vulnerabilities that exist in areas such as the microarchitecture, trusted execution environments, secure enclaves, secure boot, cryptographic instruction set architectures, memory encryption, and electromagnetic fault injection attacks. This review aims to cater to the needs of modern researchers for the development and implementation of the RISC-V ISA in a secure manner.
Authors - Nithin Kandi, Murari Nallamalli, Dorai Sai Charan M, Vijay G, Beena B. M. Abstract - Dynamic load balancing in distributed computing environments, especially with heterogeneous nodes, remains a significant challenge due to the fluctuating nature of workloads and resource availability. This paper presents a novel approach leveraging Deep Deterministic Policy Gradient (DDPG), a reinforcement learning algorithm, for optimal workload allocation in real-time systems. The system aims to minimize latency and maximize resource utilization by dynamically adapting to varying node metrics, including CPU usage, memory load, and latency. The DDPG model is trained on simulated state data, and real-time inference is performed through an API Gateway, enabling seamless integration with a five-node cluster. Results demonstrate that the proposed system outperforms traditional static and heuristic approaches in balancing workloads, optimizing resource utilization, and reducing latency. The approach is scalable, robust, and easily adaptable for edge and hybrid cloud architectures, providing a cost-effective solution for dynamic load balancing in distributed systems. This work bridges the gap between traditional cloud infrastructure and edge computing, ensuring efficient resource management in real-time systems.
Authors - Jyotiprakash Mishra, Sanjay K. Sahay, Aman Pathak Abstract - Modern processors have achieved significant performance enhancements through the implementation of speculative execution. These enhancements stem from hardware optimizations that not only improve performance but also introduce side channels, which are exploited to undermine the system’s security model. Following the discovery of Spectre, which revealed that speculative execution pipelines could bypass security boundaries, nearly all microarchitectural structures and hardware optimizations have become targets for exploitation. In the wake of these attacks, the immediate response from organizations releasing mitigation patches led to noticeable performance degradation overnight, without fully addressing the underlying issues. This paper provides a comprehensive review of architecture-agnostic attacks on modern computing systems, tracing their evolution from the initial emergence of Spectre to contemporary attacks targeting Apple Silicon. We detail the mechanisms of these attacks, the environments in which they are exploited, and their broader security implications. Furthermore, we analyze various mitigation strategies that have been proposed, acknowledging that these strategies often fail to fully resolve the issues and typically incur a performance cost. These mitigation patterns include proposed changes in operating systems, hardware, and compilers. This review aims to provide researchers and architects with the foundational knowledge needed to develop more effective mitigation strategies that address the vulnerabilities while minimizing performance overhead.
Authors - Ankit Agarwal, Ankur Pandey, Ashlesh Kumar, Dhanush D, Swetha G Abstract - Steady-State Visual Evoked Potential (SSVEP)-based Brain Computer Interfaces (BCIs) are a promising tool for non-invasive neural communication and control, particularly for individuals with severe physical or medical conditions that limit conventional interaction. However, accurately detecting and classifying SSVEP signals remains challenging due to noise and inter-subject variability. This study evaluates the performance of established classification methods, including Canonical Correlation Analysis (CCA), Filter Bank CCA (FBCCA), and transfer learning models such as EEGNet, DeepConvNet, and ShallowConvNet. To address the limitations of existing methods, we propose a novel hybrid approach combining CCA, Continuous Wavelet Transform (CWT), and Convolutional Neural Networks (CNN). This method aims to enhance feature extraction and classification accuracy. The models were evaluated on the benchmark SSVEP dataset from Tsinghua University, with preprocessing steps involving independent component analysis (ICA) and band-pass filtering. FBCCA achieved the highest accuracy of 97.5%, followed by CCA (93%) and DeepConvNet (86.95%). Our proposed method attained an accuracy of 77.52%, demonstrating its potential for robust SSVEP classification. These results underline the value of advanced algorithms and preprocessing strategies in improving SSVEP-based BCI performance, paving the way for more effective assistive technologies.
Tuesday August 25, 2026 9:30am - 11:30am IST Virtual Room AGOA, India
Authors - Radhika V. Kulkarni, Avish Agrawal, Aryan Vimal, Rohan Barde, Raghav Bajaj, Khursheed Gaddi Abstract - The Indian judicial system heavily relies on precedents for legal interpretations and decision making, providing access to relevant case law a critical yet time consuming task for legal professionals and researchers. This paper presents an AI-powered Legal Case Search Engine designed to transform legal re-search by leveraging advancements in Natural Language Processing (NLP) and Large Language Models (LLMs). The system enables efficient retrieval of contextually relevant legal precedents from the Supreme Court of India’s judgments, utilizing techniques like vector embeddings, cosine similarity, and semantic search. It offers concise case summaries and metadata insights to streamline decision making and improve accessibility to legal data. Built on open-source technology, the platform emphasizes scalability, cost efficiency, and user centric design, ensuring adaptability for future enhancements like multilingual support. By democratizing access to legal knowledge, this research aims to bridge the gap between complex legal texts and their practical application, fostering innovation in legal workflows and enhancing the rule of law.
Authors - Prajwal S, Praveen M P, Dhanya M Abstract - The global transition toward electric mobility is crucial in mitigating climate change, reducing air pollution, and promoting sustainable urban transportation. India, one of the fastest-growing markets for electric vehicles (EVs), has witnessed a surge in electric two-wheeler (E2W) adoption. However, concerns regarding battery longevity, charging infrastructure, and affordability remain key barriers to widespread adoption. This study applies sentiment analysis techniques to assess consumer perceptions of electric bikes using machine learning models for sentiment classification. A dataset comprising 3,395 consumer reviews was collected from leading automotive platforms, including BikeWale, BikeDekho, OneDrive, and ZigWheels, using web scraping techniques. The data was analyzed using VADER, TextBlob, Naïve Bayes, Logistic Regression, and Support Vector Machines (SVM) to classify sentiment and identify key consumer concerns. The results indicate a predominantly positive sentiment towards electric bikes, driven by environmental benefits and cost savings. However, consumers expressed concerns over battery efficiency, charging station availability, and high initial costs. Among the models tested, SVM achieved the highest accuracy, making it the most effective in sentiment classification. This study contributes to the limited academic research on electric bikes, offering data-driven insights into consumer perceptions. By utilizing real-world consumer data from widely used automotive platforms, the research provides valuable information for policymakers, manufacturers, and industry stakeholders. The findings aim to assist in developing strategies to address consumer concerns and enhance electric bike adoption in India.
Authors - Vani E S, Akshay Sinha, Rahul Singh Rajput, Pranjal Krishna Gupta, Chiranth K M Abstract - The primary goal of any educational institution is to provide students with a high-quality learning experience and comprehensive knowledge. Identifying students who need additional support and implementing effective strategies to enhance their academic performance is critical to achieving this objective. This study applies three machine learning techniques to develop a predictive model for assessing student performance across various academic disciplines and institutions. The techniques include Logistic Regression, k-Nearest Neighbours (KNN), and Support Vector Machine (SVM). The models were evaluated using metrics such as the Receiver Operating Characteristic (ROC) index and classification accuracy. Additional performance indicators, including classification error, precision, recall, and the F1-score were also computed. The dataset, which consists of data from a student survey and academic records, included information from a total of 700+ students. Among the models tested, the SVM model outperformed the others, achieving an ROC index of 0.82 and a classification accuracy of 84.04%.
Tuesday August 25, 2026 9:30am - 11:30am IST Virtual Room AGOA, India
Authors - Bhadouriya Khushi Mukeshsingh, Rajput Adityasingh Shashikantsingh, Parmar Smit Dharmeshkumar, Tiwari Prashant Dineshkumar, Nirav D. Mehta, Anwarul.M.Haque Abstract - Fire emergencies pose significant risks, with conventional alarms often lacking rapid response mechanisms. Delays in manual intervention can lead to severe consequences in residential and industrial settings. This study presents the Domestic Emergency System (DES), an IoT-integrated, multi-layered fire detection and emergency response framework. DES utilizes flame and smoke sensors, GSM-based emergency dialing, and RF communication modules to enable real-time hazard detection, structured evacuation, and automated alerts. An adaptive thresholding algorithm minimizes false alarms while ensuring precise fire identification. Experimental validation demonstrated fire detection within 0.8 seconds, RF signal propagation up to 25 meters, and emergency call execution within 5 seconds, improving response time by 70% compared to traditional alarms. The multi-tier alert mechanism and mobile notifications enhances situational awareness and security coordination. The DES framework outperforms conventional fire alarm systems by integrating automated emergency calling, RF-enabled intra-building communication, and a resilient multi-modal alert system. Its scalability and adaptability to smart city infrastructure make it a transformative solution for modern safety applications.
Authors - Samyuktha D, Chandrathara P T, Gowri Gireesh, Rojalin Patri Abstract - Financial literacy affects individuals' financial conduct, especially their capacity to save, consume well, and engage in thrift behaviors. In this study, the connection among financial literacy and various thrift behaviors—Careful Thrift, Spending Thrift, Saving Thrift, and Lifestyle Thrift—amongst college students is examined. Utilizing a survey quantitative methodology, statistical analysis is conducted to assess how financial literacy can predict such thrift behaviors. It also explores if demographic variables of gender and financial support sources are moderators of such relationships. Results show that financial literacy positively influences cautious thrift, lessens irresponsible spending, and increases saving and lifestyle thrift. The paper finally ends with suggested recommendations for financial literacy programs used to enhance financial decision-making skill among students.
Authors - Gauri S. Bhagat, Nitin S. More Abstract - Cardiovascular disorders continue to be a major global health issue, with arrhythmias presenting significant hurdles in both diagnosis and treatment. This article presents a groundbreaking and thorough framework for ECG feature assessment that incorporates morphological, temporal, and frequency-domain elements, all improved by advanced processing techniques and smart classification methods. By utilizing diverse features and tailoring approaches to individual patients, the proposed system enhances diagnostic accuracy and dependability. Evaluations on standard datasets indicate improved classification efficacy, marking a substantial advancement in automated systems for arrhythmia diagnosis. The framework’s adaptability further positions it as a strong prospect for integration into mobile and telemedicine platforms,thus facilitating diagnostic processes to near real-time clinical application.
Authors - Ann Mary Francis, C. Rojalin Patri, A. Varun Raj, B. Harijith M Abstract - This study evaluates the extent of customer satisfaction with private logistics service providers, where courier services are widely used. The study aims to comprehend customer perception of service quality by implementing a dual survey method: a customer service call survey and a walk-through audit. The key is to determine the areas of improvement and increase overall customer satisfaction. The findings indicate that more than 40% of the customers are dissatisfied with the service delivered. Findings indicate major problem areas are delivery experience, communication channels, and customer service interactions. The findings indicate that Order and tracking (improving order accuracy, transparency, and real-time tracking ability), Delivery (enhancing delivery timeliness, reliability, and communication during the delivery process), and Facility Factors (streamlining facility layout, staffing, and processes to facilitate smooth operations and effective service delivery) must be enhanced. Through the enhancement of these three aspects, private logistics service providers can enhance overall customer satisfaction.
Tuesday August 25, 2026 12:30pm - 2:30pm IST Virtual Room AGOA, India
Authors - Sanjeev Kumar Punia, Karanjeet Singh, Fahar Imran Abstract - The exponential growth of the World Wide Web has increased the need for efficient Web-page prediction models to reduce access latency and enhance user experience. Traditional methods, such as k-order Markov models, struggle with balancing prediction accuracy and complexity. In this work, we propose an ensemble model that combines Logistic Regression, Naive Bayes, and a First-Order Markov model to improve Web-page prediction accuracy. Logistic Regression identifies relationships in Web-log datasets, Naive Bayes applies probabilistic reasoning, and the Markov model captures transition probabilities between pages. By combining these models using a stacking classifier, we aim to leverage their strengths for more robust predictions. Our results demonstrate that the ensemble model outperforms individual models, achieving higher accuracy, precision, and recall. This hybrid approach can significantly improve Web navigation efficiency, making it a promising solution for real-time Web-page prediction.
Authors - Rupali Vairagade, Shailesh Hiralal Yadav, Harshal Raju Ismulwar, Manoj Ramashish Gupta, Nilakshi Jain, Shwetambari Borade Abstract - Digital statistics develop at an extraordinary pace, making the demand for cozy, scalable, green records more than ever before. Traditional encryption methods, including symmetric and asymmetric encryption, have long been extremely important for protected tag marks. However, these traditional methods face major challenges consisting of complex mathematical calculations, difficult intrinsic management, and obstacles to scalability. These issues can be enjoyed by improving aid, poor overall performance, and bad users. Hybrid encryption structures provide an effective solution with the help of a combination of stable security of heterogeneous encryption for critical changes and administration and the efficiency of symmetric encryption. This twin technology allows for faster processing of records, but the encryption key is protected and protected from unauthorized entries. By using the power of both encryption strategies, hybrid encryption deals with scalability and overall performance issues. This is often seen in traditional systems and is ideal for large packages in a wide range of different fields. In the long run, hybrid encryption is actually the main break in protection technology, providing a balanced solution that improves security, improves machine efficiency and ensures scalability. This approach is suitable for current desired developments for virtual environments, allowing businesses to protect sensitive statistics without compromising performance or enjoying users.
Authors - Dhruva R. Rinku, Parimi Hema Sree, D. Nagajyothi, Anita Kulkarni Abstract - In recent years, the frequency of floods has surged due to climate change and unchecked urbanization. In developing nations such as India, floods wreak havoc that can take decades to recover from. To effectively mitigate the impact of flood disasters, precise flood inundation warnings are essential. This system employs a Geographic Information System (GIS) model, constructed using a Digital Elevation Model (DEM) and building shape-files, to analyze land behavior during floods and identify inundation risks in various locations. This information is crucial for taking proactive measures to reduce flood-related destruction promptly. The Global Precipitation Measurement’s (GPM) half-hourly rain data is utilized to assess the current influence of rainfall on flood conditions across different regions, thereby enabling timely warnings to residents in nearby areas. Furthermore, this system incorporates a flood relief system, which is web-based and developed using PHP as the web application and a MySQL database. This platform prepares donors to assist flood victims with various types of donations. Immediately following the release of a flood warning through a GSM module that sends SMS alerts, donors are informed to be ready with their contributions. This timely communication equips authorities to provide refuge to flood victims. This system is specifically designed for Mumbai, India’s largest city, which is frequently affected by floods, resulting in significant property and life damage annually.
Authors - Nikita Patil, Basawaraj Patil Abstract - The goal of this paper is to leverage Field Programmable Gate Array (FPGA) technology to create a Lane Departure Warning and Correction System. Its Advanced Driver Assistance System (ADAS) design attempts to improve vehicle safety by avoiding in advertent lane changes. Under a variety of driving circumstances, such as changes in lighting, weather and road types, the system reliably detects lane boundaries by utilizing image processing techniques like Hough Transform and Canny Edge Detection. The Xilinx Zynq Ultra scale FPGA, which combines high-performance processors and peripherals for real- time processing and control, is used in the implementation. In order to guarantee that the car stays in its lane, the system is made to deal with issues like faded markers, shadows, and construction zones. It does this by promptly sending out alerts or taking remedial action. By addressing real-world issues in autonomous and semi-autonomous vehicles, this breakthrough highlights the emerging of strong hardware and cutting-edge algorithms, greatly lowering the risks associated with lane departure incidents.
Authors - Gauri Gautam, Vijay Kumar Sharma Abstract - This paper outlines an organized approach to creating a dataset and training an Artificial Neural Network (ANN) for autonomous vehicle driving systems. It utilizes the KITTI dataset and includes key preprocessing steps like converting binary files to CSV. Other steps comprise calibrating and plotting 3D point cloud data and creating 2D front-view projections with consistent sizes. A novel data preparation pipeline is presented, ensuring homogeneity while addressing challenges like variable point distributions. It also confronts cropping consistency to improve data quality. The results demonstrate a robust methodology for generating training-ready datasets. A three-layer ANN is trained effectively for autonomous driving tasks. This work contributes significantly to autonomous systems. It provides a scalable approach that can adapt to various datasets.
Authors - Jaimin Dave, Chintan Shah, Premal Patel Abstract - Implementing effective control over harmful actions in a network through an Intelligent Detection System (IDS) is necessary for modern digital security, but building robust techniques for coverage with high accuracy remains a challenge. To help overcome this challenge, the study's contribution proposes a hybrid deep learning approach combining convolution neural networks (CNN) and Long Short Term Memory (LSTM) networks for maximum coverage of detections in IDS. Tests have been conducted on various datasets and the model achieved best results of 75% detection accuracy for different attack scenarios. This was better than what's achieved using traditional methods as the legacy Intelligent Detection Systems (IDS) techniques, although increasing detection coverage, reduced the level of falsely identified cases and improved adaptability towards new patterns of attacks. The results open new frontiers for the development of hybrid machine learning architectures capable of addressing the shortcomings of traditional Intelligent Detection Systems (IDS) models and improving network security.
Authors - Safa Imtihaz Sayyad, Jyoti M Satihal, Ujwala Patil Abstract - Object detection in dark conditions at night is challenging due to poor visibility, leading to reduced accuracy and performance. We propose ODDNet, a framework for object detection in the dark that combines enhanced RGB images with event-based data, leveraging their complementary strengths. Using Zero-DCE, RGB images are enhanced for low-light, while event data is processed through a Temporal Multi-scale Aggregation to extract its temporal features. Attention mechanisms and specialized loss functions improve low-light imaging, preserve spatial details, and enhance detection accuracy. Ablation studies highlight the contributions of Zero-DCE and attention-based fusion. The proposed ODDNet model achieves a mean average precision (mAP) of 45.8% during the day and 28.3% at night at an intersection over Union (IoU) threshold of 0.5. These results demonstrate ODDNet’s capability to effectively address low-light challenges, indicating its potential for autonomous systems and night-time surveillance.
Authors - Aarti Amod Agarkar, Harshal Vijay Chaudhari, Sanskar Mukta Chaudhari, Astha Sachin Chaudhari, Om Yogesh Borse Abstract - The Personalized AI Doctor is a smart and user-friendly platform designed to make healthcare more accessible and efficient. It simplifies the process of booking medical appointments by using advanced technologies like artificial intelligence (AI), natural language processing (NLP), and secure cloud-based databases. Patients can easily register on the platform, share their symptoms through a chatbot or voice commands, and receive accurate disease predictions powered by AI. The system intelligently matches patients with the right doctors based on their specialization and availability, ensuring timely care. It also allows users to book appointments and attend virtual consultations through automatically generated Google Meet links, offering convenience and flexibility. Additionally, the platform provides detailed insights and reports for administrators, helping optimize doctor schedules and improve resource allocation. By combining AI, modern database systems, and cloud services, this system transforms the healthcare experience, making it simpler, faster, and more patient-centric.
Authors - Ganesh Puri, Pratik Jadhav, Mayuri Gawande, Gadekar Gayatri, Tushar Badakh Abstract - Air quality is a key determinant of public health, increasingly impacted by natural events such as wildfires and volcanic eruptions, as well as human-induced sources like vehicular and industrial emissions. This research focuses on predicting air pollution levels using two advanced models: Long Short-Term Memory (LSTM) networks and Auto-Regressive Integrated Moving Average (ARIMA). A dataset is developed by combining precise ground station data from multiple ground stations. It also features real-time inputs from low-cost ESP32-powered IoT sensors. The sensor fusion approach enhances both spatial coverage and data granularity, offering a more comprehensive representation of environmental conditions. The ThingSpeak IoT platform is employed for live data collection, visualization, and remote monitoring. LSTM networks are selected for their capability to model complex, long-term dependencies in time-series data, while ARIMA provides a robust statistical baseline for comparison. Experimental results demonstrate that models trained on the fused dataset significantly outperform those using individual data sources, resulting in more accurate and reliable air quality forecasts. This study presents a scalable and cost-effective framework for air quality monitoring, supporting timely, data-driven interventions by environmental agencies and contributing to improved public health management in both urban and rural areas.
Authors - Dasari Keerthi Sai Naga Sudha, Navaneeth Rajamohan, Chakravaram Hari Priya, Mandapati Bindu Sree, Beena B.M. Abstract - The present project illustrates a comprehensive crop health monitoring and recommendation system using environmental condition data in the context of optimizing agricultural practices. Crop health monitoring utilizes AWS SageMaker Studio in training and deploying a machine learning model to classify crops as either healthy or unhealthy based on environmental inputs like temperature, humidity, rainfall, N, P, K and Ph. A Flask application, developed in SageMaker Studio, is designed as the interface for a real-time crop health prediction tool, giving farmers actionable inputs for timely intervention. Finally, a crop recommendation system, using AWS SageMaker, analyzes the environmental dataset to suggest the most suitable crop for a given region. These systems combined create sustainable farming practices and contribute positively to the improvement of agriculture productivity.
Authors - Narayan Irkal, Sanjana Ambore, Praveen Bachalapur, Nihar Kulkarni, Mohammed Azharud-din, Suneeta V Budihal Abstract - This paper presents a simulation-based study of standalone 5G networks focusing on such key performance metrics as latency, throughput, and packet loss. Three network configurations-Small Cells, Massive MIMO, and Beamforming-are considered for evaluating their impact on network performance. Three types of traffic-Voice, Video, and IoT, with different data requirements-are also considered in this study. The simulation models the real-world scenario using a 3.5 GHz carrier frequency and 100 MHz bandwidth with different modulation schemes (QPSK, 16QAM, 64QAM) and coding rates (0.5, 0.7, 0.9). The performance metrics are calculated through a combination of signal-tonoise ratio, resource block allocation, and transport block size. The results are visualized by line and bar plots, emphasizing the efficiency and trade-offs of different configurations under diverse traffic scenarios. This work explains the main issues of making 5G networks more dependable and efficient based on the upsurge in demand of high-speed-low-latency in modern applications.
Authors - Zeeshan Mirji, Muzammil Kharadi, Prajwal Shiggavi, Abdul Razzak R Yergatti, Mohammed Azharuddin Adhoni, Suneeta V Budihal Abstract - Software-defined networks (SDNs) enable flexibility by decoupling the control and data planes, but their complex, dynamic structures challenge traditional optimization methods like rule-based algorithms and Queuing Theory (QT). To address this, we propose a framework using graph neural networks (GNNs) and deep reinforcement learning (DRL). GNNs model network components as nodes and connections as edges, learning efficient representations to improve metrics like delay, jitter, load balancing, and scalability. Our approach delivers scalable, real-time SDN optimization, significantly outperforming QT in simulations, paving the way for advanced data-driven network control.
Authors - Iliyas Kinnal, Yashwant Danaraddi, Amogh Gujamagadi, Aditya Deshpande, Mohammed Azharuddin, Sunita V Budihal Abstract - Analyzing the performance of the communication system in terms of real-time applications can be judged by network performance analysis. Wireless networks consist of transmission power, which determines all the key performance metrics, such as throughput, delay, packet loss, and energy efficiency. With an increase in transmission power, it can cover greater signal strength and reduce packet error but will lead to increased amount of interference, energy consumption, and potential network congestion. The NS3 network simulator provides a robust platform for modeling and analyzing the impact of transmission power on wireless networks. It will allow researches to simulate real-life scenarios that alter transmission power levels and hence the effects it will have on the performance of the network under greatly differing conditions. Critical parameters such as SNR, RSSI, and link reliability can effectively be measured through NS3, thus providing the optimum network configuration insights.
Authors - Soni R. Ragho, Rohan R. Swami, Tanuja S. Gaikwad, Aaditi P. Narke, Shubham M. Atak Abstract - India's agricultural sector is a vital pillar of its economy, providing livelihoods to millions of people. However, plant diseases position a major threat to crop productivity, leading to significant financial losses. The unpredictable nature of climate change has further exacerbated the spread of these diseases, highlighting the need for early and accurate detection to prevent large-scale crop damage. Traditional disease identification methods, which rely on human observation, are often ineffective, subjective, and prone to misdiagnosis. Incorrect assessments may result in the misuse of pesticides, causing economic burdens and environmental harm. Consequently, the development of an advanced and reliable plant disease detection system is essential for promoting sustainable farming practices. The rise of artificial intelligence and image processing has introduced innovative techniques for detecting plant diseases. Deep convolutional neural networks (CNNs) have demonstrated exceptional efficiency in identifying and classifying plant diseases with high accuracy. These models utilize sophisticated machine learning techniques to analyze high-resolution leaf images, ensuring fast and precise disease detection. This study aims to design an advanced CNN- based model to improve the accuracy and effectiveness of plant disease identification, providing farmers with actionable insights for better disease control. By harnessing AI-driven technologies, this research seeks to reduce agricultural losses, enhance crop yields, and contribute to the long-term sustainability of India’s agricultural sector. Additionally, integrating such intelligent systems can optimize resource management and reduce reliance on harmful chemical treatments.
Authors - Prerna Agrawal, Savita Gandhi Abstract - Pneumonia is a predominant cause of illness and death globally, particularly affecting vulnerable groups such as children, the elderly, and immunocompromised individuals. Timely and precise diagnosis is essential for effective therapy; however, conventional chest X-ray (CXR) evaluation by radiologists is prone to human error and constrained availability, especially in resource-limited environments. Recent breakthroughs in deep learning have facilitated automated and highly precise medical picture analysis, presenting a possible alternative for pneumonia identification. About 473,780 cases of pneumonia were reported in India in 2022–2023 and throughout this time, pneumonia was caused by 11,497 baby fatalities that aged between 1 to 12 months and 4,571 pediatric pneumonia-related deaths have been reported that aged between 1 to 5 years. In 2024 the annual incidence rate of community acquired Pneumonia in India is estimated to be between 5 and 11 per 1,000 people. This research proposes a system named PneumoSense, an intelligent automatic pneumonia diagnosis system employing DenseNet121, a deep learning model recognized for its efficacy in medical imaging applications. PneumoSense is an automated, intuitive interface that allows users to upload an x-ray image for analysis to ascertain the presence of pneumonia. Upon detection, the algorithm generates a prediction score and advises medical consultation. The model underwent comprehensive testing against various lung illnesses, such as COVID-19 pneumonia, fibrosis, and effusion, confirming its reliability in practical applications. Experimental findings indicate that DenseNet121 surpasses other deep learning algorithms, including CNN, ResNet50, and NASNet, attaining superior recall of 0.9795 and AUC score of 0.9808. The workflow of PneumoSense is also discussed in the paper. PneumoSense diminishes diagnostic inaccuracies, improves accessibility, and provides a feasible AI-driven substitute for manual diagnosis. This study underscores the revolutionary capacity of deep learning in medical diagnostics, facilitating early pneumonia detection, enhancing patient outcomes, and alleviating the workload on radiologists.
Authors - Vani K S, Aditya Andotra, Tanmay Sinha, Kaatyaini Jaiswal, Roshan Kumar Sahu Abstract - It’s very important for understanding and responding to customer emotions in real time to have an effective and satisfiable customer service. This study introduces an automated Sentiment Analysis System for Helpdesk Calls, leveraging Long Short-Term Memory (LSTM) networks and advanced Natural Language Processing (NLP) techniques to enhance service efficiency. Traditional sentiment analysis methods mostly fail to capture the nuances of spoken language, which reminds the need for a more robust approach. The proposed system processes helpdesk calls recordings, applying speech normalization, noise reduction, and feature extraction using Mel-Frequency Cepstral Coefficients (MFCCs) before classification. By integrating machine learning and deep learning models, the system provides real-time sentiment insights, allowing operators to prioritize and address calls based on emotional tone. Performance evaluation using accuracy, precision, recall, and F1-score ensures continuous model refinement. This scalable solution and methodology reduce manual effort, enhances customer interactions, and fosters improved satisfaction and loyalty, representing a significant advancement in automated customer service technology.
Authors - Pavan.B.Shivalli, Mahamadshiraj.B, Nitin.P.Savvase, Samit.Patil, Sarvesh.R.Karkannavar, Mohammed Azharuddin, Suneeta.V.Budhihal Abstract - A simulation model for inter-vehicle communication is presented in this research with the goal of improving traffic control and road safety during collisions. The model mimics data transfers between vehicles using OMNET++ and SUMO, allowing for the real-time identification and notification of traffic accidents. The framework makes use of "VEINS" to seamlessly integrate traffic and network modeling, enabling efficient data packet transfers for both routine and urgent messages. Dedicated Short-Range Communication (DSRC) with a 300-meter range and optimized power transfer at 9mW, which increases energy efficiency, are two important aspects. The framework achieves high-speed data transfer by using the UDP protocol, which is necessary for prompt response in accident situations. The findings show that vehicles can communicate reliably with one another, which could speed up emergency responses and lessen traffic interruptions caused by accidents.The model can be modified to accommodate different traffic situations and allows for the addition of more safety beacons. This model may be expanded in the future to support practical Intelligent Transportation Systems (ITS) applications.
Authors - BG.Shresta, P.Hari Sankar, Pinninti Anju Chowdary, Beena B.M Abstract - This project is to utilize cloud computing in deploying an ML-enabled Flask application for the prediction of water quality. The model, trained on historical datasets of water quality, is integrated into a user-friendly web interface constructed using Flask. Specifically, the main objective is to explore the deployment and hosting of the applications across the AWS cloud platform to ensure scalability, reliability, and efficiency. Other deployment methods scrutinized include AWS Elastic Beanstalk, AWS Lambda, and EC2. For deploying, Elastic Beanstalk was adopted as the principal deployment because it can natively host the application and handle scalability while providing end-to-end management of infrastructure. However, issues arose with regards to Lambda’s resource limitation in handling gigantic ML models, while configuration problems of EC2 were inevitable in scaling. The project itself reflects strengths and limitations of these services and points towards very robust infrastructure in resource-intensive ML applications. It’s applicable not only in the scalable solution of real-time water quality prediction but also comparative with studies on deploying Flask applications within AWS, with insights to optimize performance and future applications based on cost-effectiveness.
