Title : Performance Assessment and Prognostics of A Hybrid System using a Particle Filter approach

One challenge for the industry is to predict the performance degradation of a system based on existing observations and to take predictive maintenance action to prevent faults and failures from occurring. In this paper, a new method for hybrid system’s performance assessment is introduced, which uses Particle Filter. The novel characteristic of this prognostic method is to monitor a certain class of unknown hybrid systems by learning the system inherent behaviors from continuous observations of that system. The system’s normal or abnormal conditions are to be investigated in a discrete state space by recognizing the system’s operational state trajectory. The Rao-Blackwellised Particle Filter is employed to estimate the marginal posterior of the system’s discrete states from prior distribution. This method has been successfully demonstrated on an automatic door control system.

[1]  Neil J. Gordon,et al.  A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking , 2002, IEEE Trans. Signal Process..

[2]  D. Catlin.  Estimation, Control, and the Discrete Kalman Filter , 1988 .

[3]  Michael W. Hofbaur,et al.  Hybrid Diagnosis with Unknown Behavioral Modes , 2002 .

[4]  Nando de Freitas,et al.  The Unscented Particle Filter , 2000, NIPS.

[5]  James S. Albus,et al.  Data Storage in the Cerebellar Model Articulation Controller (CMAC) , 1975 .

[6]  Jay Lee,et al.  Machine performance monitoring and proactive maintenance in computer-integrated manufacturing: review and perspective , 1995 .

[7]  Geir Storvik,et al.  Particle filters for state-space models with the presence of unknown static parameters , 2002, IEEE Trans. Signal Process..

[8]  Nando de Freitas,et al.  Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks , 2000, UAI.

[9]  N. de Freitas,et al.  On-line probabilistic classification with particle filters , 2000, Neural Networks for Signal Processing X. Proceedings of the 2000 IEEE Signal Processing Society Workshop (Cat. No.00TH8501).

[10]  P. Fearnhead.  Markov chain Monte Carlo, Sufficient Statistics, and Particle Filters , 2002 .

[11]  V. Borkar,et al.  A unified framework for hybrid control: model and optimal control theory , 1998, IEEE Trans. Autom. Control..

[12]  M. Polycarpou,et al.  Incipient fault diagnosis of dynamical systems using online approximators , 1998, IEEE Trans. Autom. Control..

[13]  James S. Albus,et al.  New Approach to Manipulator Control: The Cerebellar Model Articulation Controller (CMAC)1 , 1975 .

[14]  Arnaud Doucet,et al.  Recursive state estimation for multiple switching models with unknown transition probabilities , 2002 .

[15]  Jun S. Liu,et al.  Sequential Monte Carlo methods for dynamic systems , 1997 .

[16]  A. Doucet,et al.  A survey of convergence results on particle ltering for practitioners , 2002 .

[17]  Michael S. Branicky,et al.  Studies in hybrid systems: modeling, analysis, and control , 1996 .

[18]  Jay Lee.  Measurement of machine performance degradation using a neural network model , 1996 .

[19]  G. Casella,et al.  Rao-Blackwellisation of sampling schemes , 1996 .

[20]  Stéphane Lafortune,et al.  Failure diagnosis of dynamic systems: an approach based on discrete event systems , 2001, Proceedings of the 2001 American Control Conference. (Cat. No.01CH37148).

[21]  Feng Zhao,et al.  FAULT ISOLATION IN HYBRID SYSTEMS COMBINING MODEL BASED DIAGNOSIS AND SIGNAL PROCESSING , 2000 .

[22]  N. D. Freitas.  Rao-Blackwellised particle filtering for fault diagnosis , 2002 .

[23]  Shengbing Jiang,et al.  Diagnosis of repeated/intermittent failures in discrete event systems , 2003, IEEE Trans. Robotics Autom..

[24]  Y. Bar-Shalom,et al.  Multiple-model estimation with variable structure , 1996, IEEE Trans. Autom. Control..

[25]  Hassane Alla,et al.  Combining hybrid Petri nets and hybrid automata , 2001, IEEE Trans. Robotics Autom..

[26]  Brian C. Williams,et al.  Mode Estimation of Probabilistic Hybrid Systems , 2002, HSCC.

[27]  Arnaud Doucet,et al.  Particle filters for state estimation of jump Markov linear systems , 2001, IEEE Trans. Signal Process..

[28]  Timothy J. Robinson,et al.  Sequential Monte Carlo Methods in Practice , 2003 .

[29]  Jitendra Tugnait,et al.  Adaptive estimation and identification for discrete systems with Markov jump parameters , 1981, 1981 20th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.

[30]  Raja Sengupta,et al.  Diagnosability of discrete-event systems , 1995, IEEE Trans. Autom. Control..