Hybrid lightweight machine learning framework for intrusion detection and mitigation in military wireless sensor networks

Abstract

Modern militaries rely heavily on Wireless Sensor Networks (WSNs) for a variety of tasks, including border security, tactical communications, drone coordination, vital infrastructure protection, and combat observation. Such networks enable secure data transmission in hostile environments while providing real-time situational awareness. However, due to limited computational resources, memory constraints, and battery dependency, WSNs are highly vulnerable to security threats such as node impersonation, Sybil attacks, replay attacks, false data injection, and message tampering. These attacks can significantly disrupt mission-critical operations and compromise sensitive military information. To address these challenges, this article proposes a Hybrid Lightweight Machine Learning-Based Intrusion Detection and Mitigation System (HL-ML-IDS) for Military Wireless Sensor Networks (MWSNs). The proposed framework combines anomaly based and signature-based detection techniques to achieve high intrusion detection accuracy with low computational overhead, making it suitable for resource-constrained military environments. The framework incorporates energy-efficient preprocessing, lightweight feature selection, adaptive threat analysis, dynamic rule generation, and an intrusion mitigation mechanism that isolates compromised nodes to prevent attack propagation. Experimental results demonstrate the effectiveness of the proposed HL-ML-IDS framework, achieving a 35% reduction in energy consumption and a 28% decrease in the false-positive rate. The framework attains a detection rate of 97.2%, precision of 96.4%, recall of 95.8%, and an F1-score of 94.9%, confirming its capability to provide secure, reliable, and energy-efficient intrusion detection and mitigation in MWSNs.