Open AccessInvariant to Observation Conditions, the Algorithm for Processing Spatially Distributed Data from a Monitoring Network Consisting of Three Sensors
Author(s) -
Елена Чернецова,
Vyacheslav Burlov,
A. Shishkin
Publication year - 2020
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/459/2/022016
Subject(s) - invariant (physics) , a priori and a posteriori , algorithm , monotonic function , false alarm , computer science , alarm , data processing , data mining , mathematics , artificial intelligence , engineering , mathematical analysis , philosophy , epistemology , mathematical physics , aerospace engineering , operating system
A multi-sensory environmental monitoring system is being considered, in which pollution data is transmitted via radio from three sensors. It is assumed that the centralized processing of information at the point of observation and control in order to detect a signal indicating the presence of contamination. The proposed processing algorithm is invariant to any transformations of data coming from sensors. It keeps the property of the data to be equally or differently distributed. Such transformations are any non-monotonic functions of the original observations. It is shown that with respect to these transformations, the maximum invariant statistics is the set of ranks of quantities taken as elements of a single sample. The developed detection algorithm has a sufficiently high resistance to observation conditions, since it retains its working capacity when the observation conditions are non-stationary. The algorithm is also invariant to the law of fluctuations of the pollution signal and its location among the last two observed sites and provides a constant probability of false alarm in any (a priori unknown) noise distribution.