Transforming Urban Traffic with AI: Insights from Singapore and Opportunities in India
Real-time traffic management has become the backbone of modern urban planning, with AI systems at the forefront of optimizing traffic flow and reducing congestion. With urbanization fast-paced worldwide, cities face unprecedented traffic challenges, such as increasing vehicle density, unpredictable congestion patterns, and growing environmental concerns. This paper reviews the AI-based traffic management system implemented in Singapore, a global leader in smart city innovation. Advanced techniques of AI are put forward by Singapore's LTA in managing traffic such that intelligent traffic lights come through predictive analytics, amongst its integration with public transportation means, cutting significant congestion, travel time as well as vehicle emissions levels alongside improving road safety all across. The potential applicability of such systems is also discussed in Indian cities like Mumbai and Bengaluru. With high population density, diverse traffic compositions, and infrastructure constraints, cities are more demanding and require innovative solutions to handle these challenges. AI-based traffic management could be applied to adjust the timing of traffic signals dynamically, optimize public transport efficiency, and reduce emergency response times for transformative changes in urban mobility. However, fragmented data systems, infrastructure limitations, and cost barriers pose huge implementation challenges in India. By comparing Singapore's success to the realities of Indian cities, this research highlights what's needed to adapt and scale AI technologies to meet local needs. It concludes that the integration of AI-driven systems can provide Indian cities with a sustainable path forward regarding traffic congestion, reduced environmental impact, and the quality of life in an urban setting.
Published by: Satheerth P K
Author: Satheerth P K
Paper ID: V10I6-1471
Paper Status: published
Published: December 25, 2024
AI-Based Cancer Detection using FRCNN, Random Forest, SVM, and Regression Models
Early detection of cancers, especially in the mouth, throat, and lungs, significantly improves patient survival rates. This paper presents a comprehensive AI-driven approach combining various machine learning (ML) and deep learning (DL) techniques such as Fast Region-based Convolutional Neural Networks (FRCNN), Random Forest (RF), Support Vector Machines (SVM), and Logistic and Linear Regression models to enhance cancer detection capabilities. Each model's strengths are leveraged to create a hybrid system that excels in detecting and classifying cancerous regions in medical images and analyzing patient data. The proposed workflow incorporates automated image analysis, feature selection, classification, and probabilistic risk estimation, enhancing diagnostic accuracy while addressing challenges like data availability, model interpretability, and computational requirements. This integrated AI-based approach demonstrates potential for real-time clinical application and personalized cancer diagnostics.
Published by: Shivam Chattar, Anshul Gaikwad, Het Savsani, Kshanay Nikam, Avishkar Sarnaik, Prof. Priyanka Patil
Author: Shivam Chattar
Paper ID: V10I6-1450
Paper Status: published
Published: December 23, 2024
Sustainable Information Retrieval Techniques for Onion Market Instability Prediction using Machine Learning and Deep Learning Approaches
Price is a critical determinant in financial activities, and sudden fluctuations in price often signal market instability. Machine learning offers robust techniques to forecast product prices and address these instabilities effectively. This study examines the application of machine learning models to predict onion prices in India, utilizing data collected from the Ministry of Agriculture, India. Various machine learning algorithms, including K-Nearest Neighbor (KNN), Naïve Bayes, Decision Tree, Neural Network (NN), and Support Vector Machine (SVM), were employed for classification purposes. Their performance was evaluated to identify the most accurate model. Additionally, this research integrates deep learning approaches, specifically Long Short-Term Memory (LSTM) networks, for forecasting onion prices. These methods classify onion prices into three categories: preferable (low), economical (mid), and expensive (high), providing valuable insights to address market volatility.
Published by: Dr. M.K. Jayanthi Kannan, Satyajit Patel
Author: Dr. M.K. Jayanthi Kannan
Paper ID: V10I6-1455
Paper Status: published
Published: December 22, 2024
A Support Vector Regression Model for Air Quality Prediction in Lucknow
Air quality significantly impacts public health, particularly in urban areas like Lucknow, where deteriorating air quality has been linked to severe health issues, especially in children and vulnerable groups. Accurate air quality prediction allows authorities to implement timely measures to shield these populations from harmful exposure. A lack of comprehensive data and robust algorithms has limited traditional forecasting methods. This study employs a Support Vector Regression (SVR) model to forecast pollutant levels and the Air Quality Index (AQI) in Lucknow using publicly available historical data from the Central Pollution Control Board (CPCB) and local monitoring stations. Among various configurations, the SVR model with a Radial Basis Function (RBF) kernel showed superior performance, achieving an accuracy of approximately 93.4%. Utilizing all available variables rather than relying on feature selection methods like Principal Component Analysis (PCA) improved prediction outcomes. The model effectively forecasts key pollutants, including sulfur dioxide (SO2), carbon monoxide (CO), nitrogen dioxide (NO2), particulate matter (PM2.5 and PM10), and ground-level ozone (O3). This research demonstrates the potential of advanced machine learning techniques to address air quality challenges in Lucknow, offering valuable insights for policymaking and urban environmental management.
Published by: Vivek Chauhan, Rahul kumar, Dr. Nidhi Saxena
Author: Vivek Chauhan
Paper ID: V10I6-1425
Paper Status: published
Published: December 21, 2024
FastMonitor: Enhancing Data Access Control with Zero-Trust Architecture
As organizations grapple with the escalation of data breaches and sophisticated cyber threats, ensuring robust and adaptive data access control has become a top priority. Traditional perimeter-centric security frameworks, once the backbone of enterprise protections, have increasingly shown limitations against evolving adversarial tactics. In response, Zero-Trust Architecture (ZTA) has emerged as a dynamic paradigm, rejecting implicit trust and demanding continuous verification of every access request. This article introduces FastMonitor, a data-access control solution designed to operationalize ZTA principles to their full potential. Introduced in June 2022, FastMonitor rigorously authenticates and authorizes each data access request, consolidating granular policies, comprehensive audit trails, and real-time monitoring into a single coherent system. By bridging theoretical concepts of ZTA with practical enforcement mechanisms, FastMonitor significantly mitigates unauthorized access, supports compliance with stringent regulatory frameworks, and strengthens organizational resilience. This study elucidates the architectural design of FastMonitor, demonstrates its capacity to enhance security in sectors as diverse as healthcare and automotive, and examines its role in fostering greater national cybersecurity readiness.
Published by: Frank Mensah
Author: Frank Mensah
Paper ID: V10I6-1453
Paper Status: published
Published: December 21, 2024
Zero Trust Architecture: A Comprehensive Review of Principles, Implementation Strategies, and Future Directions in Enterprise Cybersecurity
In an era characterized by digital transformation and increasingly sophisticated cyber threats, traditional perimeter-based security models have become inadequate for safeguarding modern enterprise IT infrastructures. Zero Trust Architecture (ZTA) emerges as a pivotal paradigm shift, fundamentally redefining organizational cybersecurity by eliminating implicit trust and enforcing continuous verification of every access request. This review paper provides an in-depth examination of ZTA, tracing its evolution from foundational principles articulated by Forrester Research and the National Institute of Standards and Technology (NIST) to its contemporary extensions addressing the complexities of diverse and decentralized digital environments. Key components of ZTA, including context-aware and continuous authentication, device authentication, and robust encryption mechanisms, are meticulously analyzed to elucidate their roles in enhancing security posture. The paper also explores the logical architecture of ZTA, highlighting the interplay between Policy Engine, Policy Administrator, and Policy Enforcement Points, which collectively enforce stringent access controls and monitor ongoing activities. Despite its advantages, the implementation of ZTA presents significant challenges, such as integration with legacy systems, operational overhead, and vulnerabilities related to policy decision processes and insider threats. Best practices for successful ZTA adoption are discussed, emphasizing comprehensive asset inventory, strong identity and access management, micro-segmentation, continuous monitoring, and phased implementation approaches. Furthermore, the review identifies emerging trends and future directions, including the integration of ZTA with 5G networks, Internet of Things (IoT), edge computing, artificial intelligence, machine learning, post-quantum cryptography, and blockchain technology. By synthesizing insights from recent studies and industry frameworks, this paper aims to provide a holistic understanding of Zero Trust Architecture, offering valuable guidance for organizations seeking to enhance their cybersecurity resilience in an ever-evolving digital landscape.
Published by: Frank Mensah
Author: Frank Mensah
Paper ID: V10I6-1452
Paper Status: published
Published: December 21, 2024
