DQN-Based Anti-Jamming Resource Scheduling for Unmanned Aerial Vehicle Networks with MF-TDMA

In this paper, a MAC layer anti-interference strategy based on deep Q-network (DQN) is proposed for clustered unmanned aerial vehicle network (UAVNET) operating in dynamic interference environment. The network is divided into multiple clusters, each of which is managed by a cluster head responsible for intra-cluster and inter-cluster communication. The cluster head is a deep Q-network (DQN) agent that dynamically allocates frequency and selects transmission links to maximize system throughput while mitigating the impact of directional interference attacks. The network is built based on long short-term memory (LSTM) to capture temporal dependencies. To further address the long-term fairness issue between destination nodes, an α-fairness mechanism is integrated in the reward function. Simulations under different interference probabilities ($p=0.3,0.5,0.7,0.9$) show that the proposed method outperforms the traditional static MF-TDMA scheme and improves the throughput by $28 \%$ in high interference conditions. In addition, the introduction of α-fairness makes the distribution of service opportunities between drones more balanced while maintaining robust network performance. These results verify the effectiveness and adaptability of the proposed DQN-based MAC strategy in complex drone communication environments.