Enhanced Active Eavesdroppers Detection System for Multihop WSNs in Tactical IoT Applications

In tactical Internet of Things (IoT) applications, the broadcast nature of wireless sensor networks (WSNs) makes it easy to eavesdrop on their traffic. Furthermore, adversaries can access the sensor nodes to intercept and eavesdrop on critical wireless transmission. Research in this area focuses on reducing the eavesdropping probability through specific methods, such as encryption and transmission power control. However, eavesdropper detection techniques in WSNs do not exist in the literature. This article proposes a novel enhanced active eavesdroppers detection (EAED) system for homogeneous multihop WSNs. The EAED system consists of a monitoring module and a detection engine module. The Monitoring module plays a vital role in the EAED system to provide accurate measurements for the detection engine module. We propose three monitoring architectures for this measurement: 1) static monitoring nodes; 2) unmanned aerial vehicles (UAVs)-based monitoring; and 3) neighborhood monitoring. To find the optimal locations for static monitoring nodes, we use a genetic algorithm (GA). We also use the Hamiltonian path planning to calculate the flight path for UAVs. The detection engine module utilizes a lightweight anomaly detection method that employs the
$Z$
-test method and runs on edge devices. We analyze and discuss the network overhead, advantages, and disadvantages of different monitoring architectures. According to the simulation results, the EAED system can detect active eavesdroppers with a high acrlong DR
$({\ge }90\%)$
and a low false-positive rate
$({\leq }5\%$
) and outstanding performance (
${\mathrm{ AUC}}\approx 0.97$
).