A Hidden Markov Model Application with Gaussian Mixture Emissions for Fault Detection and Diagnosis on a Simulated AUV Platform

摘要

This paper presents an application of a Hidden Markov Model for fault detection and diagnosis on a testbed that emulates an AUV thruster system. The testbed consists in circuit board with two DC motors that represent the thrusters and embedded features to produce malfunctions. We present how the model is learned using the Expectation Maximization algorithm for Gaussian Mixtures and how the testbed is monitored probabilistic inference. Diagnosis is also performed using GMM classifiers. We describe how the framework deals with non-Gaussian data and how it reflects in the accuracy overall.