Authomatic impulse noises filtration for measurement systems based on the Bayesian statistics

I.E. Tarasov

The increase in complexity of data-measuring systems puts in higher claims to the functionality of independent components. One of the tendencies in instrument-making is the extension of the list of the tasks performed by a primary measuring device being a data source, and in this connection such term is used as “an intelligent sensor”. The version combining a sensing element and a computing device, which performs the primary information processing, permits to formulate the task of the search of new methods and algorithms of this processing capable of effective utilization of increasing computing possibilities of digital microelectronic devices. The article deals with the features of the statistical method based on Bayes theorem of improved noise immunity under moderate-volume analyzed samples compared to arithmetic mean estimation. Its usage in primary handlers of computing information seems to be perspective as it allows shifting a portion of computing load to sensors and digital data processing devices combined with them.