Hidden Markov Models and Neural Networks for Fault Detection

摘要

Continuous online monitoring of complex dynamic systems is common in applications as diverse as industrial plant operations, telecommunications network, and biomedical health monitoring. For monitoring purposes, the exact nature of the system under observation is typically not relevant provided there exist some measurements or symptoms which provide diagnostics information regarding the underlying system state.