Title :
Extensions of jump Markov system estimation techniques
Author :
Sun, Kunpeng ; Li, Lun ; Panayotopoulos, George ; Packard, Andy
Author_Institution :
Dept. of Mech. Eng., California Univ., Berkeley, CA, USA
Abstract :
In this paper, we review the framework for fault modeling and identification using jump Markov systems. We derive a special case of an estimation problem where the Viterbi algorithm holds, allowing the most-likely model path to be computed recursively in time, with fixed computational requirements at each time step. In the same case, the model probability at each time step can also be calculated exactly, with fixed computation. We also present a new heuristic, and evaluate it´s performance on a well-known example.
Keywords :
Markov processes; fault location; maximum likelihood estimation; Viterbi algorithm; fault identification; fault modeling; jump Markov system estimation; model probability; Control systems; Fault detection; Fault diagnosis; Mechanical engineering; Probability; Random processes; Random variables; Recursive estimation; Sun; Viterbi algorithm;
Conference_Titel :
American Control Conference, 2002. Proceedings of the 2002
Print_ISBN :
0-7803-7298-0
DOI :
10.1109/ACC.2002.1024592