This paper explores the adaptation of Federated Learning (FL) to enhance cybersecurity in industr... more This paper explores the adaptation of Federated Learning (FL) to enhance cybersecurity in industrial control systems, which are vital for critical infrastructures. Integrate ICS into digital networks to better control distant or remote places, which increases sophistication in cyber-attacks, thereby threatening traditional centralized IDS solutions for ICS. Centralized systems are inefficient for ICS's needs as they adopt a decentralized and privacy-preserving aspect. FL is an approach to decentralized machine learning that promises solutions in the realm of multiple entities training security models on their data while hiding sensitive operational information. This study evaluates the potential applicability of FL in solving some of the issues ICS faces under increasing cyber-attacks and draws a comparative analysis with traditional IDS versus FL-based approaches. Findings will assist in the betterment of ICS security, thereby ensuring the smooth operation of critical infrastructures against increasingly sophisticated cyber threats.
World Journal of Advanced Research and Reviews, 2023
This paper discusses about U.S. electrical grid which is a vital infrastructure supporting a lot ... more This paper discusses about U.S. electrical grid which is a vital infrastructure supporting a lot of industries and people around the USA. However, it faces various challenges because of aging components, the threat of extreme weather, and increasing energy demands. Due to this, it will become extremely difficult to maintain the resilience and reliability of the grid and a growing concern for utilities, policymakers, and stakeholders. By using the quantitative method, the research shows potential benefits including a 20% reduction in unplanned outages, with a 15% improvement in operational efficiency that is supported by a 20% reduction in unplanned outages and just 15% improvement observed in operational efficiency level, supported by cost-benefit analysis. This research explores in detail the potential of machine learning, predictive analytics, and Internet of Things (IoT) sensors to modernize the electrical grid, minimize downtime caused by component failure, and enhance efficiency. Therefore, by implementing the historical data, advance machine learning models, real-time data monitoring, and predictive maintenance is helpful to identify main failures present in critical components like transmission lines, transformers, and substations before they occur. This study investigates in detail the design and implementation of a predictive analytics platform reliable for the U.S. grid by focusing on machine learning algorithms, data collection, and scalability challenges. The findings focus on the need for strategic collaboration between policymakers, utilities, and technology providers to minimize challenges related to data integration, cost, and infrastructure. This research is contributing to the ongoing efforts for building a highly resilient and sustainable electrical grid, capable of meeting the required demand of the future and minimizing risks caused by aging infrastructure.
Global Journal of Engineering and Technology Advances, 2025
This report investigates the implementation of advanced machine learning models within Supervisor... more This report investigates the implementation of advanced machine learning models within Supervisory Control and Data Acquisition (SCADA) systems to enhance intrusion detection capabilities and system security. By utilizing models such as CatBoost and XGBRegressor, which excel in processing complex, non-linear data, the study demonstrates significant improvements in predicting and managing operational states in wind turbines. The incorporation of Explainable AI (XAI) techniques, particularly SHAP values, further provides transparency in model decisions, fostering trust among stakeholders. Recommendations are provided for effective model integration, deployment with XAI features, and necessary policy enhancements to ensure the secure, reliable, and ethical use of AI in critical infrastructure environments.
International Journal of Computer Applications Technology and Research, 2022
The advent of quantum computing poses an existential threat to contemporary cryptographic standar... more The advent of quantum computing poses an existential threat to contemporary cryptographic standards, particularly those securing decentralized blockchain networks and cloud infrastructures. Classical public-key cryptosystems such as RSA, ECC, and DH, which rely on factorization and discrete logarithm problems, are rendered obsolete by Shor's algorithm, necessitating the transition toward post-quantum cryptographic (PQC) solutions. This study explores the integration of PQC algorithms, including lattice-based, hash-based, code-based, multivariate, and isogeny-based cryptographic mechanisms, within blockchain-ledger technologies and cloud architectures to ensure long-term security against quantum adversaries. A comparative analysis is conducted to evaluate computational efficiency, key size implications, communication overhead, and security resilience under quantum attack models. The research highlights the adaptation of PQC within blockchain consensus mechanisms, smart contract execution, and cryptographic primitives such as digital signatures, zero-knowledge proofs, and secure multi-party computation (MPC). Additionally, it examines the impact of PQC on cloud security, addressing challenges in quantum-safe key exchange protocols, homomorphic encryption for secure computations, and cross-platform interoperability within hybrid quantum-classical cloud ecosystems. Real-world implementations and benchmarking data provide insights into the feasibility of large-scale adoption, shedding light on standardization efforts by NIST and industry consortia. The study concludes with future directions, emphasizing the need for efficient PQC algorithm optimization, lightweight cryptographic frameworks for IoT-driven blockchain applications, and scalable post-quantum identity management systems. By establishing quantum-resistant security frameworks, this research underscores the imperative need for early adoption to mitigate cryptographic vulnerabilities in the impending post-quantum era.
The growing complexity and scale of Internet of Things (IoT) ecosystems demand robust and interpr... more The growing complexity and scale of Internet of Things (IoT) ecosystems demand robust and interpretable security solutions to detect and mitigate emerging threats. This research integrates Explainable AI (XAI) techniques, including SHAP (SHapley Additive Explanations) and LIME (Local Interpretable Model-Agnostic Explanations), into machine learning models such as Random Forest and SVM to enhance transparency and trust in IoT threat detection systems. A distributed processing architecture, utilizing Apache Spark and Kafka, ensures real-time scalability for processing large volumes of heterogeneous IoT data. XAI methods provided actionable insights by identifying key features influencing predictions, significantly reducing false positives and improving analyst response times. The study aligns with regulatory requirements like GDPR, offering a scalable, interpretable framework for IoT security. However, challenges such as computational overheads and adversarial risks highlight areas for future research.
Uploads
Papers by smart idima