Sampling in Factored Dynamic Systems

Koller, Daphne; Lerner, Uri

doi:10.1007/978-1-4757-3437-9_21

Cited by 35 publications

(47 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PFs have been used extensively for centralized system health monitoring and diagnosis applications [7,12]. Distributed inference schemes, such as the BK algorithm [15], creates individual factors by eliminating causal links between weakly interacting subsystems.…”

Section: Discussionmentioning

confidence: 99%

“…The system parameters are deterministic. A PF is a sequential Monte Carlo sampling method for Bayesian filtering that approximates the belief state of a system using a weighted set of samples, or particles [7]. Each sample, or particle, consists of a value for each state variable, and describes a possible system state.…”

Section: Distributed Diagnosis Approachmentioning

confidence: 99%

“…Each DBN-F and BG-F pair is used to derive a local diagnoser. We implement a particle filter (PF)-based inference approach [7] on each DBN-F for fault detection, isolation and identification, and use the BG-F for qualitative fault isolation. By construction, the random variables in each DBN-F are guaranteed to be conditionally independent of the random variables in all other DBN-Fs, given a subset of communicated measurements considered as system inputs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Using factored bond graphs for distributed diagnosis of physical systems

Roychoudhury

Biswas

Koutsoukos

2010

Proceedings of the 2010 Spring Simulation Multiconference

View full text Add to dashboard Cite

This paper presents a distributed Bayesian fault diagnosis scheme for physical systems. Our diagnoser design is based on a procedure for factoring the global system bond graph (BG) into a set of structurally observable bond graph factors (BG-Fs). Each BG-F is systematically translated into a corresponding DBN Factor (DBN-F), which is then used in its corresponding local diagnoser for quantitative fault detection, isolation, and identification. By construction, the random variables in each DBN-F are conditionally independent of the random variables in all other DBN-Fs, given a subset of communicated measurements considered as system inputs. Each DBN-F and BG-F pair is used to derive a local diagnoser that generates globally correct diagnosis results by local analysis. Together, the local diagnosers diagnose all single faults of interest in the system. We demonstrate on an electrical system how our distributed diagnosis scheme is computationally more efficient than its centralized counterpart, but without compromising the accuracy of the diagnosis results.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Distributed Diagnosis Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Using factored bond graphs for distributed diagnosis of physical systems

Roychoudhury

Biswas

Koutsoukos

2010

Proceedings of the 2010 Spring Simulation Multiconference

View full text Add to dashboard Cite

show abstract

“…This approach can be employed for large-scale systems with components having all types of failure distributions (the one applied is Weibull's). Koller and Lerner (2000) elicited dynamic Bayesian networks for monitoring dynamic systems. It is pointed out that hidden Markov model processes and Kalman filters are particular cases of DBNs.…”

Section: Bayesian Network and Fault Diagnosismentioning

confidence: 99%

Bayesian reliability models of Weibull systems: State of the art

Zaïdi

Bouamama

Tagina

2012

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

In the reliability modeling field, we sometimes encounter systems with uncertain structures, and the use of fault trees and reliability diagrams is not possible. To overcome this problem, Bayesian approaches offer a considerable efficiency in this context. This paper introduces recent contributions in the field of reliability modeling with the Bayesian network approach. Bayesian reliability models are applied to systems with Weibull distribution of failure. To achieve the formulation of the reliability model, Bayesian estimation of Weibull parameters and the model's goodness-of-fit are evoked. The advantages of this modelling approach are presented in the case of systems with an unknown reliability structure, those with a common cause of failures and redundant ones. Finally, we raise the issue of the use of BNs in the fault diagnosis area.

show abstract

“…However, autonomous transitions between modes triggered by the continuous dynamics have not been considered. Particle filtering has been applied also for a class of hybrid systems modeled by dynamic Bayesian networks in [7] where the autonomous transitions between discrete states are defined using the so-called softmax conditional probability distributions. A fault modeling and diagnosis approach for hybrid systems based on qualitative representation of the fault hypotheses has been presented in [8].…”

Section: Introductionmentioning

confidence: 99%

Estimation of Distributed Hybrid Systems Using Particle Filtering Methods

Koutsoukos

Kurien

Zhao

2003

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Networked embedded systems are composed of a large number of components that interact with the physical world via a set of sensors and actuators, have their own computational capabilities, and communicate with each other via a wired or wireless network. Such systems are best modeled by distributed hybrid systems that capture the interaction between the physical and computational components. Monitoring and diagnosis of any dynamical system depend crucially on the ability to estimate the system state given the observations. Estimation for distributed hybrid systems is particularly challenging because it requires keeping track of multiple models and the transitions between them. This paper presents a particle filtering based estimation algorithm for a class of distributed hybrid systems. The hybrid estimation methodology is demonstrated on a cryogenic propulsion system.

show abstract

Sampling in Factored Dynamic Systems

Cited by 35 publications

References 0 publications

Using factored bond graphs for distributed diagnosis of physical systems

Using factored bond graphs for distributed diagnosis of physical systems

Bayesian reliability models of Weibull systems: State of the art

Estimation of Distributed Hybrid Systems Using Particle Filtering Methods

Contact Info

Product

Resources

About