Vol. 8 No. 1
Year: 2013
Issue : Jul-Sep
Title : Performance Evaluation Of Fuzzy Voters Used In Safety Critical Systems
Author Name : Phani Kumar, SeethaRamaiah Panchumarthy , Anu Gokhale
Synopsis :
The main objective of this research paper is to design a self configurable dynamic fuzzy voter that can be used in the NModular Redundant (NMR) systems to mask the fault and exhibit better safety performance. Proposed dynamic voter can configure itself for the changing data of any range and changing deviations in the redundant module outputs due to the changes in the operational modes of the safety- critical embedded systems while in operation. In this paper, existing fuzzy voters are extensively surveyed and the merits and demerits or limitations are discussed. The major limitation observed in the existing fuzzy voters is the preselected optimal static fuzzy parameter selection technique. Preselected optimal fuzzy parameters may work for only specific ranges of data and specific amount of deviations among the redundant modules outputs of NMR systems but fail for the changing data ranges and deviations conditions. A dynamic fuzzy parameter selection method using statistical parameters on the local data in each voting cycle is proposed in this paper and applied to the existing static fuzzy voters. Static and dynamic versions of the fuzzy voters are empirically evaluated for 10,000 voting cycles and safety performance is plotted. Dynamic voters have shown improved safety performance compared to their static counterparts and are more useful for the changing operational modes and conditions.
For more details
No comments:
Post a Comment