Risk assessment for autonomous navigation system based on knowledge graph

This paper presents a knowledge graph construction framework based on the Adversarial Interpretive Structure Model (AISM), aiming to reveal the failure interdependencies of autonomous navigation assemblies. The autonomous navigation system is divided into 6 systems with 43 failure modes. The hierarchical correlations between multiple systems are quantified by establishing the distance matrix and adjacency matrix, which are defined as triples to serve the entities and attributes for the knowledge graph. The centrality metrics are applied to evaluate the importance of nodes and to assess the stability of the knowledge graph. The coordination relationships between autonomous navigation systems and the impact associations of failure modes are discussed to validate the reasonability of the proposed framework. The results indicate that The NAVTEX operation panel and The installation base of the temperature sensor are the most essential components in the hierarchical results. The positioning system has a high degree centrality and betweenness centrality, while the server system has the highest closeness centrality. The propagation path from Decrease in positioning accuracy to The installation base of the temperature sensor is investigated, providing the methodological and data basis for decision-making.