From Static to Adaptive Defense: Federated Multi-Agent Deep Reinforcement Learning-Driven Moving Target Defense Against DoS Attacks in UAV Swarm Networks

TL;DR

This paper tackles DoS threats in low-altitude UAV swarm networks by shifting from static defenses to an adaptive, distributed strategy. It proposes a federated multi-agent deep reinforcement learning framework (PG-FMADRL) that coordinates three lightweight moving target defense actions—leader switching, route mutation, and frequency hopping—under a multi-agent POMDP, with parameter aggregation at a central aggregator to balance generalization and per-agent customization. Empirical results show substantial improvements: attack mitigation rates up to 0.999, recovery times reduced by up to 94.6%, and notable reductions in energy usage (up to 29.3%) and cumulative defense costs (up to 98.3%) across various DoS strategies, including greedy attackers. The framework demonstrates strong resilience and scalability, suggesting practical potential for secure, reliable, and energy-efficient low-altitude networks, while also highlighting future work on decentralized aggregation and co-evolving adversaries.

Abstract

The proliferation of UAVs has enabled a wide range of mission-critical applications and is becoming a cornerstone of low-altitude networks, supporting smart cities, emergency response, and more. However, the open wireless environment, dynamic topology, and resource constraints of UAVs expose low-altitude networks to severe DoS threats. Traditional defense approaches, which rely on fixed configurations or centralized decision-making, cannot effectively respond to the rapidly changing conditions in UAV swarm environments. To address these challenges, we propose a novel federated multi-agent deep reinforcement learning (FMADRL)-driven moving target defense (MTD) framework for proactive DoS mitigation in low-altitude networks. Specifically, we design lightweight and coordinated MTD mechanisms, including leader switching, route mutation, and frequency hopping, to disrupt attacker efforts and enhance network resilience. The defense problem is formulated as a multi-agent partially observable Markov decision process, capturing the uncertain nature of UAV swarms under attack. Each UAV is equipped with a policy agent that autonomously selects MTD actions based on partial observations and local experiences. By employing a policy gradient-based algorithm, UAVs collaboratively optimize their policies via reward-weighted aggregation. Extensive simulations demonstrate that our approach significantly outperforms state-of-the-art baselines, achieving up to a 34.6% improvement in attack mitigation rate, a reduction in average recovery time of up to 94.6%, and decreases in energy consumption and defense cost by as much as 29.3% and 98.3%, respectively, under various DoS attack strategies. These results highlight the potential of intelligent, distributed defense mechanisms to protect low-altitude networks, paving the way for reliable and scalable low-altitude economy.

Figures (7)

• Figure 1: Bird's-eye view of the proposed DoS mitigation approach.
• Figure 2: Overview of the proposed lightweight and coordinated Moving Target Defense (MTD) mechanisms for UAV swarm networks. (a) Leader Switching: Dynamically reassigns the logical leader role among UAVs, thereby enhancing resilience by preventing persistent targeting of a single node/link; (b) Route Mutation: Alters communication paths by selecting alternative relay UAVs, enabling critical messages to bypass compromised links and maintain connectivity under attack; (c) Frequency Hopping: Simultaneously changes the communication frequency channels of all UAVs and the ground control station, increasing uncertainty for attackers and disrupting their ability to sustain effective DoS attacks.
• Figure 3: The average return of the proposed PG-FMADRL method under different DoS attack types and strategies.
• Figure 4: The recovery time of the UAV swarm under different solutions for defeating DoS attacks with different types and strategies.
• Figure 5: The energy consumption of the UAV swarm under different solutions for defeating DoS attacks with different types and strategies.