Federated Multi-Agent Deep Reinforcement Learning-Driven Moving Target Defense Against DoS Attacks in UAV Swarm Networks

View a PDF file from the paper entitled Fixed Defense to Air Conditioning: Modern Modern Agents targeted the target defense to learn against DOS attacks in drone squadron networks, pen Cho and 5 other authors

PDF HTML (experimental) view

a summary:The spread of drones enabled a wide range of important applications and has become the cornerstone of low height networks, smart cities support, emergency response, and more. However, the open wireless environment, dynamic topology, and resource restrictions from drones expose low -rise networks to severe DOS threats. Traditional defense methods, which depend on fixed configurations or central decision -making, cannot respond effectively to the quickly changing conditions in the drone swarm environments. To face these challenges, we suggest a new framework for multi-agent learning (FMADRL)-the target defense framework (MTD) driven by DOS in low height networks. Specifically, we design lightweight and coordinated MTD mechanisms, including switching a leader, road boom, and frequency navigation, to disable the attacker’s efforts and enhance network flexibility. The defense problem is formulated as a multi -agent Markov decision that can be partially observed, capturing the unconfirmed nature of drones under attack. Each drone is equipped with a political factor that independently chooses MTD procedures based on partial notes and local experiences. By employing a political gradient algorithm, drones work to improve their policies cooperative through the weighted assembly of bonuses. Intensive simulation operations show that our approach greatly exceeds the modern foundation lines, achieving up to 34.6 % improved in the rate of mitigation of the attack, reducing the average recovery time up to 94.6 %, a decrease in energy consumption and the defense cost of up to 29.3 % and 98.3 %, respectively, within the various DOS attack strategies. These results highlight the potential of smart defense mechanisms distributed to protect low rise networks, paving the way for the reliable and low -development economy.