As a core information infrastructure in the 6G era, the Space-Air-Ground Integrated Network (SAGIN) integrates space-based, air-based, and ground-based network resources to achieve seamless communication across all domains. However, its characteristics such as heterogeneous node coupling and dynamic topology changes make it prone to cascading failures, severely threatening critical business continuity in Internet of Things (IoT) applications spanning smart cities, healthcare, transportation, and industrial automation. This paper conducts systematic research addressing challenges including modeling difficulties in SAGIN cascading failure propagation, insufficient coordination of defense strategies, and poor resource adaptability. First, a multi-factor coupled dynamic model of cascading failure propagation is established to quantify the synergistic effects of node heterogeneity, link dynamics, and load redistribution. Second, a closed-loop collaborative defense system integrating “early warning-isolation-self-healing” is designed. The system incorporates a lightweight greedy-based self-healing algorithm and uses multi-criteria decision-making (Analytic Hierarchy Process) for resource optimization. These approaches ensure real-time performance and energy efficiency on resource-constrained edge nodes. Third, a joint simulation platform combining NS-3 and MATLAB is built to validate the model and strategies across diverse IoT application scenarios. Experimental results show that the proposed propagation model maintains prediction error within 10%, the defense strategies increase failure recovery rates to 85%–90%, reduce communication interruption duration by over 60%, and lower resource overhead by 20%–25%, providing theoretical support and technical guarantees for stable SAGIN operation in security and resiliency-critical environments.

To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising from modal heterogeneity during fusion, while also capturing shared information across modalities, this paper proposes a Multi-modal Pre-synergistic Entity Alignment model based on Cross-modal Mutual Information Strategy Optimization (MPSEA). The model first employs independent encoders to process multi-modal features, including text, images, and numerical values. Next, a multi-modal pre-synergistic fusion mechanism integrates graph structural and visual modal features into the textual modality as preparatory information. This pre-fusion strategy enables unified perception of heterogeneous modalities at the model’s initial stage, reducing discrepancies during the fusion process. Finally, using cross-modal deep perception reinforcement learning, the model achieves adaptive multi-level feature fusion between modalities, supporting learning more effective alignment strategies. Extensive experiments on multiple public datasets show that the MPSEA method achieves gains of up to 7% in Hits@1 and 8.2% in MRR on the FBDB15K dataset, and up to 9.1% in Hits@1 and 7.7% in MRR on the FBYG15K dataset, compared to existing state-of-the-art methods. These results confirm the effectiveness of the proposed model.