P. Lázaro, R. Barco, and J.M. Hermoso (Spain)
earth station, bayesian network,symptom, hypothesis, ranging, tracking
The intrinsic real-time nature of space missions demands prompt response from the ground segment side to any incidents that may arise during 24h daily operations, especially to performance anomalies. This fact, together with the increasing complexity in the ground segment structure, makes it quite hard, even for the experienced human operator, to accurately diagnose the ultimate cause of the problem in order to accordingly take further action. The development of a tool which can perform these diagnosis tasks either autonomously or as an aid element to the operator appears therefore to be essential. In this scenario, this paper describes the basic principles of Bayesian networks and the performance of an earth station, and introduces the novel concept of applying Bayesian networks to real-time fault isolation in earth stations. An example of the Bayesian network constructed is presented, along with the development of a tool for automated diagnosis of earth stations based on Bayesian networks.
Important Links:
Go Back
