Multivariate multichannel software and measurement complex for detecting anomalous states of natural and technical objects and systems

An approach to the multivariate classification of the states of natural-technical objects and systems is considered, based on the development of methods for dynamic detection of anomalies in information data flows. The approach is based on an estimate of the statistical discrepancy between the probability distributions of random variables over variably changeable time intervals, as well as an estimate of the probabilities of errors of the first and second kind. The structure of a multichannel software and measurement complex for detecting anomalous states of PTO and PTS is proposed, and the results of model calculations are presented. The use of the multivariate approach allows optimizing the processes of processing, analysis and integration of heterogeneous data, as well as increasing the sensitivity, reliability and efficiency of decisions.

Markov model of data measurement complex for track geometry car

A stochastic model of a nonadjustable data measurement complex platform for track geometry cars is introduced. A state vector evaluation algorithm based on the approximation of a posteriori probability density by the system of a posteriori moments is also offered.