The Rate Gyroscope Fault Detection and Diagnose Based on ARTMAP Network
Li Guan · Ceshi jishu xuebao · 2001
Aim To introduce a method for the fault detection and diagnose of rate gyroscope. Metods Through the statistic test in sliding window, Fuzzy ARTMAP neural network is adopted to classify model parameters directly. Results This method can diagnose the fault correctly not only for the known fault but also for the unknown fault of rate gyroscope. Conclusion The method in this paper avoids the insufficient of other forward networks in sorting, It has the ability of detecing unknown faults and has no necessary to get te physical parameters of system.