Anti In-Band Full-Duplex Interception for Wireless Tactical Networks by Joint Passive Defense, Active Defense, and AES-Based Encryption

Ensuring secure and reliable communication remains a major challenge for Tactical Vehicular Ad Hoc Networks (TVANs), particularly in maintaining both security and quality of service (QoS). State-of-the-art security solutions for TVANs rely on a single anti-interception technique, often compromising QoS to counter in-band full-duplex eavesdroppers. To address this limitation, we propose a triple-layered anti-interception strategy that jointly combines passive defense, active defense, and AES-based encryption implemented at the physical layer. We formulate a non-convex optimization problem that jointly optimizes jamming allocation (JA), power allocation (PA), and key length assignment (KA). This problem is approximated via a first-order Taylor expansion and solved iteratively using a difference-of-convex (DC) approach. To reduce computational complexity under high-mobility ground combat conditions, we reformulate the problem using a Multi-Agent Deep Reinforcement Learning (MADRL) framework. Simulation results show that our MADRL-based approach closely approximates the optimal solution in near real-time and significantly outperforms state-of-the-art baselines in both Low Probability Interception (LPI) and QoS.