Statistical Algorithms for Analysis, Measurement, and Recognition of Transport Noises

In connection with the problem of geoecology of transport noise for urban infrastructure, this paper considers the detection and recognition of transport types based on seismic oscillations recorded at the ground surface and acoustic oscillations recorded in the atmosphere. A distinctive feature of this problem is that it is focused on solutions in low and infra-low frequency regions. Railway, automobile, heavy wheeled and tracked transport are considered as objects under study. The detection of oscillation sources is based on an adaptive threshold algorithm using the maximum likelihood criterion. In selected sections of records, informative features of the transport types being considered are determined by using spectral and spectral-time analysis. Based on the resulting statistics of the records, the problem of constructing etalons is solved at the stage of training and recognition of transport objects by noise in conditions of a priori uncertainty regarding their number. Numerical estimates of the accuracy of solving the problem in a field experiment are obtained.