E. Smirnova ALGEBRA OF LIMITED PETRI TS-NET IN SIMULATION TASKS OF INTERACTIVE INFORMATION SYSTEMS WITH HYPERTEXT STRUCTURE

N.R.Allahverdiyeva, A.M.Mehdiyeva CORRECTING FILTRATION OF METRICAL SIGNALS

On the modern stage paramount value is conducted optimization of processes of exchange (getting, transformations, collection, issues, processing and presentations) metrical information in information-measurement systems (IMS) for raising accuracy of results of measurements, checking, diagnosticses and object identifications to automations with provision for available technical-econmic restrictions.

At optimization IMS usually on the first plan is brought forth problem to minimization an inaccuracy of measurements. Decision of this problem is vastly complicated under numerical measurements of unceasing and changing during values. However exactly such output signals characteristic of the big class of the sources information. The Second obstacle to optimization IMS is a deficit a priori information for characteristics, measured values, disturbing factors, intermediate and end results of deciding the main functions and problems, as well as submachines an metrical and computing subsystems. At the syntheses optimum IMS, working in the dynamic mode, problem of negotiation a priori uncertainty else more is aggravated. Considering that high accuratical dynamic measurements become mass, given problem actual for the further development information-mertrical technology (technicians) as a whole.

Correcting filtration of rezults of measurements have goodly value for creation more effective communicatory metrical systems.

In practice this problem solves in two directions:

-a shaping and entering a correction in the result steady-state (nearly statical) measurements;

- correcting filtration of results of dynamic measurements.

Significant results in the direction of raising accuracy of steady-state measurements received in functioning (working) the many researchers. However, using these results to dynamic measurements possible under essential restrictions on the frequency spectrum of signals.

Majority of computing algorithms of secondary processing, analysis and presentation of results an user has a single-line nature, so they sensitive to noises, being present in all input metrical data. This circumstance speaks of that that correcting filtration it is necessary and in step of numerical analog value measurements, and in step of the preprocessing of metrical data.

In work by us made syntheses optimum high-frequency and low-frequency correcting filters for raising accuracy of metrical tract for the broad class assessed on the input signal of noises with provision for the deficit a priori information of comparatively the most useful signal.

We have also made the syntheses of optimum algorithms of processing metrical data with noise with provision for filterring characteristics of used processing operators.