By Topic

Maximum Likelihood Failure Diagnosis in Finite State Machines Under Unreliable Observations

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

3 Author(s)
Eleftheria Athanasopoulou ; Coordinated Science Laboratory, and the Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, IL, USA ; Lingxi Li ; Christoforos N. Hadjicostis

In this paper, we develop a probabilistic methodology for failure diagnosis in finite state machines based on a sequence of unreliable observations. Given prior knowledge of the input probability distribution but without actual knowledge of the applied input sequence, the core problem we consider is to choose from a pool of known, deterministic finite state machines (FSMs) the one that most likely matches the given sequence of observations. The problem becomes challenging because of sensor failures which may corrupt the observed sequence by inserting, deleting, and transposing symbols with certain probabilities (that are assumed known). We propose an efficient recursive algorithm for obtaining the most likely underlying FSM, given the possibly erroneous observed sequence. The proposed algorithm essentially allows us to perform online maximum likelihood failure diagnosis and is applicable to more general settings where one is required to choose the most likely underlying hidden Markov model (HMM) based on a sequence of observations that may get corrupted with known probabilities. The algorithm generalizes existing recursive algorithms for likelihood calculation in HMMs by allowing loops in the associated trellis diagram. We illustrate the proposed methodology using an example of diagnosis in the context of communication protocols.

Published in:

IEEE Transactions on Automatic Control  (Volume:55 ,  Issue: 3 )