Resilience analysis and critical node identification for unmanned system of systems based on multi-layer flow betweenness and adjacency information entropy

Unmanned system of systems (USoS) is a collection of multiple unmanned systems, characterized by functional dependencies and structural vulnerabilities arising from its constituent heterogeneous entities. In multi-perturbation environments, the failure of critical nodes will compromise USoS resilience and precipitate performance degradation. However, traditional node centrality methods struggle to capture the functional and coupling features of USoS, which impair the effect of protection strategies. This study presents a novel critical node identification method and resilience enhancement framework to strengthen USoS resistance and recovery capability. Firstly, a functional dependency network (FDN) model was established to characterize the node heterogeneity and operation loop structure. Secondly, a novel critical node identification method is proposed to identify an information hub in both intra-layer and inter-layer of FDN. Thirdly, a two-phase resilience evaluation method was developed based on resilience factors to consider mission phases and baseline. Additionally, a hybrid resilience enhancement strategy is proposed to bolster USoS performance by integrating pre-attack rational resource allocation with post-attack mission reallocation. Finally, several cases involving 90-nodes are illustrated to validate the efficacy of these approaches. The proposed framework achieved resilience improvement by 21.88% to 64.82%, which provides support for the resilient design and resource allocation of USoS in multi-perturbation scenarios.