Machine Learning-Driven Optimization and Security in IoT Ubiquitous Sensor Networks: A Comprehensive Review

Abstract: Internet of Things Ubiquitous Sensor Networks (IoT-USNs) represent a significant paradigm shift in technologicalinfrastructure, enabling the development of intelligent applications across various sectors, including healthcare, smart urban settings, precision agriculture, and industrial automation. Notwithstanding their widespread implementation, IoT-USNs face enduring challenges that hinder their scalability, operational efficiency, and security. This systematic review investigates cutting-edge solutions through an exhaustive analysis of 19 peer-reviewed articles published from 2018 to 2024. The identified critical limitations encompass an increased vulnerability to cyber threats, severe constraints in energy and computational resources, and complexities associated with the management of real-time data via effective aggregation and routing mechanisms. The decentralized architecture exacerbates challenges pertaining to data integrity and security enforcement, rendering networks susceptible to various attack vectors. The analysis delineates three primary research trajectories: optimization techniques driven by machine learning that achieve energy efficiency improvements of up to 40%; trust management systems that enhance authentication protocols; and adaptive routing protocols that significantly mitigate congestion issues. This review provides a critical assessment of existing methodologies, identifies notable deficiencies in integrated security-energy optimization frameworks, and underscores the limited real-world implementation of crosslayer solutions. The contribution of this work lies in synthesizing contemporary research trajectories and proposing future research directions that emphasize integrative strategies incorporating advanced security mechanisms, energy-aware protocols, and intelligent data management frameworks to fully harness the potential of IoT-USNs.