Journal of Science Engineering Technology and Management Science

E.PRAVEEN1 YVNM SARMA 2The Advanced Peripheral Bus (APB) is a low-power, low-bandwidth bus that is part of the Advanced Microcontroller Bus Architecture (AMBA) developed by ARM. It is widely used to connect low-speed peripheral devices such as timers, UARTs, GPIOs, SPI, and I²C controllers in System-on-Chip (SoC) designs. Since APB is responsible for communication between the processor and peripheral devices, verifying its functionality is essential to ensure reliable data transfer and protocol compliance. This project presents the verification of the APB protocol using the Universal Verification Methodology (UVM). UVM is a standardized verification framework built on SystemVerilog that enables the development of reusable, scalable, and modular verification environments. The project implements a complete UVM-based testbench consisting of transaction objects, sequences, sequencer, driver, monitor, agent, environment, scoreboard, and functional coverage components to verify the APB design. The verification environment generates both directed and randomized read/write transactions to validate the APB protocol under different operating conditions. The monitor captures the DUT activity, while the scoreboard compares the expected and actual results to identify mismatches. Functional coverage is collected to ensure that all important protocol features, including read operations, write operations, valid address ranges, wait states, and transfer types, are exercised during simulation. Simulation results confirm that the APB design behaves according to the protocol specification, with successful verification of all functional scenarios. The UVM-based environment improves verification efficiency, enhances code reusability, and provides comprehensive functional coverage. This project demonstrates an industry-standard verification approach for AMBA APB and serves as a strong foundation for verifying more complex bus protocols such as AHB and AXI.

Keywords : The rapid growth of System-on-Chip (SoC) technology has led to the integration of multiple processors, memories, and peripheral devices onto a single chip.

Author : E.PRAVEEN1 YVNM SARMA 2

Title : A REVIEW ON ADVANCED PERIPHERAL BUS INTERFACE PROTOCOL

Volume/Issue : 2026;03(07)

Page No : 717-724

The rapid advancement of Large Language Models (LLMs) has significantly transformed artificial intelligence by enabling machines to perform complex natural language understanding, reasoning, content generation, and decision-making tasks. Despite their remarkable performance, existing LLM-based agents primarily rely on static pre-trained knowledge, making them incapable of continuously adapting to newly emerging information and userspecific experiences. This limitation often results in outdated knowledge, limited personalization, contextual inconsistency, and catastrophic forgetting when models are retrained. To overcome these challenges, this research proposes a comprehensive roadmap for Lifelong Learning of Large Language Model Based Agents, integrating continuous knowledge acquisition, persistent memory, adaptive reasoning, Retrieval-Augmented Generation (RAG), vector-based semantic memory, knowledge graphs, explainable artificial intelligence (XAI), and feedback-driven learning mechanisms. The proposed framework enables intelligent agents to acquire knowledge incrementally from user interactions, external repositories, and dynamic information sources while preserving previously learned knowledge through efficient memory consolidation techniques. A semantic vector database stores contextual representations that facilitate fast retrieval and context-aware reasoning during future interactions. Furthermore, knowledge graph integration enhances structured knowledge representation and semantic relationship discovery, thereby improving inference capability and decision-making accuracy. Explainable AI modules increase system transparency by providing interpretable reasoning pathways, while continuous feedback learning refines knowledge quality and response generation through iterative optimization. An analytics dashboard continuously monitors memory growth, retrieval efficiency, reasoning performance, and learning progression, enabling proactive system evaluation and improvement. Experimental evaluation demonstrates that the proposed lifelong learning framework significantly improves contextual continuity, factual consistency, adaptive learning capability, response relevance, and user personalization compared with conventional LLM architectures. Moreover, the framework effectively reduces hallucination, enhances knowledge retention, minimizes redundant retraining, and supports scalable deployment across diverse real-world applications including education, healthcare, enterprise knowledge management, scientific research, software engineering, and intelligent virtual assistants. The proposed roadmap establishes a robust foundation for developing next-generation autonomous AI agents capable of continuous learning, long-term knowledge evolution, and intelligent decision-making in dynamic environments.

Keywords : Large Language Models, Lifelong Learning, Continual Learning, Retrieval-Augmented Generation, Knowledge Graphs, Vector Database, Explainable Artificial Intelligence, Intelligent Agents.

Author : 1ANTHATI KARTHIK, 2A. MANDHAN CHARI

Title : LIFELONG LEARNING OF LARGE LANGUAGE MODEL BASED AGENTS: A ROADMAP

Volume/Issue : 2026;03(07)

Page No : 725-735

Blockchain technology has become a transformative platform for secure digital transactions by enabling decentralization, transparency, immutability, and distributed trust across multiple application domains. Despite these advantages, existing blockchain platforms primarily rely on conventional cryptographic algorithms such as RSA and Elliptic Curve Cryptography (ECC), which are vulnerable to emerging quantum computing attacks. Furthermore, modern blockchain ecosystems face increasing cybersecurity challenges, including fraudulent transactions, malicious smart contracts, unauthorized access, identity spoofing, network intrusions, and sophisticated cyberattacks that cannot be effectively addressed using traditional security mechanisms alone. To overcome these limitations, this paper proposes a Quantum Secure Blockchain Framework with Machine Learning-Based Threat Detection and Smart Contract Validation, which integrates post-quantum cryptography, intelligent threat detection, trust evaluation, and blockchain security within a unified architecture. The proposed framework employs quantum-resistant cryptographic algorithms such as CRYSTALS-Kyber, CRYSTALSDilithium, Falcon, and SPHINCS+ to secure authentication, digital signatures, and transaction verification against future quantum attacks. Simultaneously, machine learning models, including Random Forest and anomaly detection techniques, continuously analyse transaction behaviour, network traffic, and user activities to identify fraudulent operations and predict security risks before transaction confirmation. A trust management engine computes dynamic trust scores based on historical transaction behaviour, while smart contract validation mechanisms verify transaction authenticity, authorization privileges, fraud probability, and compliance before blockchain execution. The framework is implemented using Python, Django, Solidity, Web3.py, Scikit-Learn, Ethereum, and MySQL, providing secure communication between decentralized applications and blockchain networks. Experimental analysis demonstrates that the proposed architecture significantly improves fraud detection accuracy, enhances transaction integrity, strengthens smart contract security, reduces malicious activities, and provides long-term resilience against both classical and quantum cyber threats. Consequently, the proposed framework establishes a scalable, intelligent, and future-proof blockchain ecosystem suitable for secure financial systems, healthcare, digital identity management, supply chain applications, and next-generation decentralized services.

Keywords : Blockchain Security, Post-Quantum Cryptography, Machine Learning, Threat Detection, Smart Contract Validation, CRYSTALS-Kyber, CRYSTALS-Dilithium, Falcon, SPHINCS+, Random Forest, Ethereum, Web3.py, Fraud Detection, Trust Management, Quantum Computing.

Author : 1D SAI BHANU VARSHIT, 2K UMA

Title : QUANTUM SECURE BLOCKCHAIN FRAMEWORK WITH MACHINE LEARNING-BASED THREAT DETECTION AND SMART CONTRACT VALIDATION

Volume/Issue : 2026;03(07)

Page No : 736-747

Autonomous systems have become a fundamental component of modern intelligent technologies owing to their capability to perform complex tasks with minimal or no human intervention. These systems are extensively employed in robotics, autonomous vehicles, unmanned aerial vehicles, industrial automation, healthcare, intelligent transportation, and smart manufacturing. However, conventional autonomous systems primarily depend on predefined rules and static control strategies, making them less effective in dynamic and uncertain environments. Reinforcement Learning (RL) has emerged as a promising machine learning paradigm that enables autonomous agents to learn optimal behaviours through continuous interaction with their surroundings. Unlike supervised learning, RL does not require labelled datasets; instead, it learns by maximising cumulative rewards obtained through trial-and-error exploration. This research presents a Reinforcement Learning-based Autonomous System implemented using a custom GridWorld environment integrated with the Deep Q-Network (DQN) algorithm. The proposed framework enables an autonomous agent to observe environmental states, execute actions, receive reward feedback, and continuously refine its policy to achieve optimal decision-making. A Flaskbased web application is developed to provide an interactive interface for user authentication, training management, model evaluation, dashboard analytics, and reinforcement learning model storage. The environment incorporates obstacles, target locations, and navigation paths that simulate real-world autonomous navigation scenarios. The DQN algorithm combines deep neural networks with Q-learning to estimate optimal action values and improve learning efficiency within high-dimensional state spaces. Experimental evaluation demonstrates that the autonomous agent gradually improves navigation efficiency, obstacle avoidance capability, cumulative reward, and policy convergence through repeated training episodes. The framework provides real-time monitoring of training performance, reward progression, and model evaluation metrics, thereby enhancing transparency and usability. The proposed system offers a scalable, adaptive, and intelligent platform that significantly improves autonomous decision-making while reducing dependence on manually programmed rules. Furthermore, the framework establishes a strong foundation for future developments involving multi-agent reinforcement learning, explainable artificial intelligence, autonomous robotics, intelligent transportation systems, and real-world adaptive control applications.

Keywords : Reinforcement Learning, Autonomous Systems, Deep Q-Network, Deep Reinforcement Learning, GridWorld, Artificial Intelligence, Autonomous Navigation, Intelligent Decision Making, Flask Framework, Policy Optimisation.

Author : 1KURSAM SNEHA, 2V. VANI

Title : REINFORCEMENT LEARNING FOR AUTONOMOUS SYSTEM

Volume/Issue : 2026;03(07)

Page No : 748-757

The rapid expansion of the Internet of Things (IoT) has revolutionized the way smart devices communicate, collaborate, and exchange information across diverse application domains such as healthcare, industrial automation, transportation, agriculture, smart homes, and smart cities. However, the increasing number of interconnected devices has significantly expanded the attack surface, making secure and reliable device authentication one of the most critical challenges in IoT environments. Conventional authentication mechanisms primarily rely on passwords, cryptographic keys, certificates, and centralized identity management systems. Although these methods provide basic protection, they are unable to effectively detect compromised devices, insider attacks, zero-day threats, or dynamic behavioural changes after authentication. Moreover, supervised machine learning approaches require extensive labelled datasets, which are expensive to obtain and often fail to generalize to emerging attack patterns. To overcome these limitations, this paper proposes a Self-Supervised Deep Learning Framework for Secure and Efficient IoT Device Authentication based on behavioural analysis and anomaly detection. The proposed framework employs a Deep Autoencoder trained using unlabeled telemetry data collected from IoT devices, including CPU utilisation, memory usage, packet transmission rate, communication frequency, and sensor activity. By learning normal operational behaviour, the model identifies abnormal activities through reconstruction error without requiring labelled attack data. A dynamic trust management mechanism evaluates each device according to its behavioural consistency and anomaly score, enabling continuous authentication rather than one-time credential verification. Devices with high trust scores are granted network access, whereas suspicious devices are isolated to prevent malicious activities. The framework integrates telemetry collection, anomaly detection, trust evaluation, authentication decision-making, and real-time dashboard monitoring into a unified security architecture. Experimental evaluation demonstrates that the proposed framework improves authentication accuracy, enhances anomaly detection capability, reduces false alarms, minimises computational overhead, and provides adaptive protection against evolving cyber threats. Consequently, the proposed approach offers a scalable, intelligent, and resource-efficient security solution suitable for next-generation IoT ecosystems requiring autonomous and continuous device authentication.

Keywords : Internet of Things (IoT), Self-Supervised Learning, Deep Autoencoder, Device Authentication, Anomaly Detection, Trust Management, Deep Learning, Cybersecurity, Telemetry Analysis, Continuous Authentication.

Author : 1G. LAVANYA, 2RAJITHA KATHULA

Title : SELF-SUPERVISED DEEP LEARNING FRAMEWORK FOR SECURE AND EFFICIENT IoT DEVICE AUTHENTICATION

Volume/Issue : 2026;03(07)

Page No : 758-769

The rapid expansion of urban populations, increasing vehicle ownership, and growing transportation demands have intensified challenges such as traffic congestion, road accidents, environmental pollution, fuel wastage, and inefficient route planning within modern smart cities. Conventional traffic management systems rely heavily on fixed traffic signal schedules, manual surveillance, and historical traffic analysis, making them inadequate for responding to dynamic and real-time transportation conditions. Artificial Intelligence (AI) has emerged as a transformative technology capable of enhancing Intelligent Transport Systems (ITS) by enabling predictive analytics, automated traffic monitoring, intelligent decision-making, and adaptive route optimisation. This study presents an AI-driven Intelligent Transport System that integrates Machine Learning, Computer Vision, and graphbased routing algorithms to improve transportation efficiency and urban mobility. The proposed framework employs a Random Forest classifier to predict traffic congestion by analysing historical traffic records, weather conditions, temporal features, and traffic density patterns. Computer Vision techniques based on YOLOv8 are utilised for real-time vehicle detection, while the MediaPipe framework performs pedestrian detection and crowd density estimation using surveillance images and video streams. Furthermore, Dijkstra's shortest path algorithm is implemented to determine optimal travel routes by considering congestion conditions and road network characteristics. The entire framework is deployed through a Flask-based web application integrated with an SQLite database and an interactive dashboard for real-time transportation analytics and decision support. Experimental evaluation demonstrates that the proposed framework achieves high prediction accuracy, reliable object detection, and efficient route optimisation, thereby reducing travel time and improving traffic flow. The integration of Artificial Intelligence technologies enables proactive traffic management, enhances road safety, supports emergency response planning, and contributes to sustainable urban transportation. The proposed Intelligent Transport System provides a scalable and cost-effective solution for smart city administrators, transportation authorities, and urban planners seeking intelligent, data-driven transportation management capable of addressing the evolving mobility requirements of future smart cities.

Keywords : Artificial Intelligence, Intelligent Transport Systems, Smart Cities, Machine Learning, Random Forest, YOLOv8, MediaPipe, Traffic Congestion Prediction, Vehicle Detection, Route Optimisation, Dijkstra Algorithm, Urban Mobility.

Author : 1CHILUKA SANDHYA RANI, 2VARUGU. RAMESH BABU

Title : THE ROLE OF ARTIFICIAL INTELLIGENCE IN INTELLIGENT TRANSPORT SYSTEMS FOR SMART CITIES

Volume/Issue : 2026;03(07)

Page No : 770-780

Privileged Access Management (PAM) systems play a critical role in securing high-risk identities within enterprise environments. Current PAM solutions, however, are mostly based on static policies and context-independent monitoring, which are not enough to detect context-dependent anomalous behaviors in contemporary distributed systems. In this paper, a context-aware, AI-based anomaly detection mechanism based on hybrid deep learning architectures to learn temporal and behavioral patterns of privileged access activities is proposed. The proposed CNN-BiLSTM-Attention network, Transformer-GRU fusion network, Attention-CNN-LSTM network and Hierarchical BiGRU with context gating network are integrated together. Experimental results on a synthetic enterprise PAM dataset show the remarkable detection performance of the Hierarchical BiGRU with an accuracy of 99.14% and an F1 score of 99.32%. The work extends PAM from a static-to-dynamic approach to access control to a more continuous approach to security monitoring.

Keywords : Privileged Access Management, Identity and Access Management, anomaly detection, deep learning, contextual risk scoring, explainable AI, enterprise cybersecurity.

Author : Anupam Nandan

Title : AI-DRIVEN PRIVILEGED ACCESS MANAGEMENT (PAM): A CONTEXT-AWARE DEEP LEARNING FRAMEWORK FOR PREDICTIVE ANOMALY DETECTION IN ENTERPRISE SYSTEMS

Volume/Issue : 2026;03(07)

Page No : 781-790