DDQN Based Adaptive Multi-Channel MAC Protocol for UAV Ad-Hoc Networks

Unmanned aerial vehicle ad-hoc networks (UANETs) are widely used in military and civilian applications. This paper introduces a double deep Q-Networks (DDQN) based adaptive multi-channel media access control protocol (DAM-MAC) for UANETs. The proposed protocol enables each node to maintain an available channel table to select suitable channels and sets different contention windows for each channel through the DDQN algorithm, dynamically adjusting based on historical information and current network conditions. A multi-channel simulation model on the OMNeT++ platform is employed to emulate the complex communication environment in UANETs. The simulation result shows that DAM-MAC significantly improves network performance such as packet loss, latency and throughput compared to the Q-Learning MAC protocol.