Analysis of the amplitude distributions of the hydro acoustic river ambient noise using ARMAsel algorithm
Miodrag Vračar, MSc (Eng)[1])
In this work ARMAsel algorithm was used for automatic estimation of the ARMA parameters of the river hydro acoustics ambient noise to evaluate the nature of the river noise amplitude distributions. Namely, on the base of the estimated ARMA parameters digital filter was realized. This filter is used for numeric synthesis of the artificial signals on the bases of various types of numerical distributions, such as: exponential, double exponential, normal, Bernoulli and uniform. The bispectrums were calculated for all of those signals. The bispectrums of artificial signals were compared with bispectrum of the ambient noise signal using two methods and results were systematized in the form of table review. This procedure enabled finding out which bispectrum of artificial signals is closest to the bispectrum of the ambient noise signal. Time duration of all the analyzed signals was in the interval from 100 to 1365 ms.
Key words: hydroacoustics, ambient noise, ARMAsel algorithm, bispectrum.