WEATHER THREAT ASSESSMENT BASED ON DYNAMIC BAYESIAN NETWORK

This paper aims to ascertain the variation in weather threat over time. To this end, the author established an assessment model of weather threat based on the Dynamic Bayesian Networks (DBN). Then, the Hidden Markov Model (HMM) was introduced to assess the level of weather threat. To validate the proposed method, a simulation was carried out on an Unmanned Aerial Vehicle (UAV), considering all five key weather factors affecting UAV safety. The results show that the DBN-based model is much more effective than the ordinary Bayesian Network (BN) in handling fuzzy information and complex weather conditions. The research findings shed new light on the accurate rout planning of an UAV.