Intelligent Air Traffic Management System Based on Knowledge Graph

In recent years, there have been many studies on knowledge graphs but few studies in air management systems. Nowadays, the problem of insufficient deployment capacity like flight delay, occurs in existing civil air traffic control systems in China. The authors aim to construct the knowledge graph of air traffic control system and design related intelligent applications. This research supports some application scenarios such as flight delay analysis and so on. The authors propose a knowledge embedding model Translating-spatio-temporal Embedding (TransST), supporting spatiotemporal information, which is evolved from the Translating Embedding (TransE) model after incorporating the embedding of information on time and space. The knowledge triples are mapped onto the hyperplane determined by the spatiotemporal information. The experiment is based on the real dataset and an encyclopedic dataset from YAGO. It shows that knowledge inference effect of Trans-ST is better than that of traditional model and state-of-art model.