Software fault diagnosis model of AUV based on Bayesian Networks and its simplified method

According to the uncertainty broadly existed in fault diagnosis of AUV software system, this paper presents a Bayesian Networks diagnosis model with three layers based on CME. On the basis of that, the paper also presents a cutting irrelative node method based on task context according to AUV´s specific nature, this method predigests network, reduces the complexity of consequence calculation, and enhances the ability of real-time fault diagnosis effectively.