Modeling and Analysis of Security Connectivity in UAV Swarm Network: An N+1-Player Zero-Sum Game

This paper investigates the modeling and analysis of attack and defense interactions in a UAV swarm network, where a dynamically flying UAV swarm suffers from attacks launched by a smart attacker, resulting in the interception and disconnection of UAVs from the swarm. First, to evaluate the security performance of a single UAV and the security connectivity of the UAV swarm, the concept of security entropy is defined and derived based on insecure probability and degree-based and connected component-based weights. Following the security connectivity model, a N+1-player zero-sum game model is formulated, where the attacker and each UAV aim to minimize and maximize the security entropy of the UAV swarm by optimizing the allocations of attack and defense resources, respectively. The existence of Nash equilibrium (NE) is proved and closed-form expressions for the NE are derived, following which a modified water-filling algorithm is proposed to obtain solutions of the NE. Numerical results reveal different resource allocation strategies of resource-dominant and resource-dominated players. Moreover, the security entropy of the UAV swarm improves with the enhanced communication connectivity of the UAV swarm.