SARRA: A Secure Adaptive and Resilient Routing Framework of FANETs with Hybrid Blockchain and Deep Q-Network Intelligence

Unmanned Aerial Vehicles (UAVs), referred to as Flying Ad Hoc Networks (FANETs), play a crucial role in disaster response due to their ability to enable rapid communication without the need for existing infrastructure. However, their high mobility, frequent link disruptions, energy problems, and security risks all have a detrimental effect on the networks' performance and stability. To address these challenges, this paper introduces SARRA (Secure Adaptive Resilient Routing and Authentication), a hybrid intelligent model. SARRA integrates Deep Q-Network-based routing, lightweight blockchain authentication, Particle Swarm Optimization for efficient cluster head selection, Ant Colony Optimization for robust path recovery, and a Federated Learning-based Intrusion Detection System to detect threats while preserving privacy. The DQN agent provides active choices of the best paths to forward by considering multidimensional state variables such as residual energy, the quality of links, hop distance, degree of congestion, and the trust score, and blockchain provides decentralized and tamper-proof node authentication. Decades of simulations on the NS-3 environment demonstrate that the SARRA is superior to standard AODV, LEACH, and base DQN protocols in terms of Packet Delivery Ratio improvement of 19.6 percent, end-to-end delay reduction by 17.3 percent, average energy consumption decrease by 22.8 percent, and throughput increase by 25.1 percent. In addition, In addition, 18.4 more routing overhead is minimized, the network lifetime is prolonged by 21.7 and the FL-IDS detects intrusions with an accuracy of 94.6 with a false positive probability of 3.2. The experiment outcome validates that SARRA is an effective solution to provide highly resilient and fault-tolerant disasterresponse networking by virtue of the fact that it greatly increases the security, adaptability, and fault tolerance of FANET communication systems.