Tracking and restrictability in discrete event dynamic systems
SIAM Journal on Control and Optimization
Why Event Observation: Observability Revisited
Discrete Event Dynamic Systems
Fault Detection and Diagnosis in Distributed Systems: An Approach by Partially Stochastic Petri Nets
Discrete Event Dynamic Systems
State Observation and Diagnosis of Discrete-Event SystemsDescribed by Stochastic Automata
Discrete Event Dynamic Systems
State Estimation of λ-free Labeled Petri Nets with Contact-Free Nondeterministic Transitions*
Discrete Event Dynamic Systems
Global state estimates for distributed systems
FMOODS'11/FORTE'11 Proceedings of the joint 13th IFIP WG 6.1 and 30th IFIP WG 6.1 international conference on Formal techniques for distributed systems
| Hi-index | 0.00 |
Knowledge of the current system state is crucial to many discrete event systems (DESs) applications such as control, diagnosis and prognosis. Due to limited sensing capabilities, the current state information is generally not available and needs to be estimated. In this paper, we propose a novel distributed state estimation algorithm for discrete event plants. According to the proposed algorithm, local sites maintain and update local state estimates based on their local observations of the plant behavior and the observations of the plant behavior sent from the other sites over communication channels with delays. For efficiency of storage, redundant history information about the possible plant evolution is truncated each time a local state estimate is updated. At each local site, the truncation is performed independently requiring no synchronization among the sites. The state estimate maintained at each of the local sites is shown to remain finite regardless of whether the system can execute an unbounded sequence of unobservable events. It is also shown that the proposed algorithm is sound and complete, i.e., each local estimate always contains the true current states (soundness), and it only contains the reachable states of the traces which give rise to a same history of observations (as received from the plant and the other local sites) as does the one executed by the plant (completeness). Also the proposed algorithm can support an architecture in which there is no communication from a certain site to certain other sites. An illustrative example is provided to demonstrate the proposed distributed state estimation algorithm.