An intelligent real-time intrusion detection and prevention strategy for emergency response MANETs

Abstract In recent years, Mobile Ad Hoc Networks (MANETs) have emerged as a pivotal domain within wireless technologies. Emergency Response MANETs (ER-MANETs) support communication among emergency responders during critical scenarios. Ensuring the real-time integrity and availability of these highly dynamic networks requires robust security measures beyond traditional approaches. This article introduces a real-time Intelligent Intrusion Detection and Prevention System (IIDPS) tailored for ER-MANETs to detect and mitigate five primary Denial of Service (DoS) attacks: Black hole, Gray hole, Flooding, Scheduling, and Traffic Interception. The proposed framework integrates Support Vector Machine (SVM) for attack classification, Random Forest Classifier (RFC) for feature selection, and Bayesian Optimization with Gaussian Process for hyper-parameter tuning, coordinated via a Central Network Administrator Unit (CNAU). Simulated across dynamic terrain topologies in NS-3, the system achieves a 97.5% classification accuracy and reduces total energy consumption by 60% to 80% under severe attack conditions. Furthermore, scalability evaluations across 50 to 200 nodes confirm low response latency (18.7 to 28.6 ms) and sustained Packet Delivery Ratio (91.8%), demonstrating high operational effectiveness for emergency deployments.