An estimation theory of level distribution over a long time interval on the basis of the level distribution over a short time interval under a digital measurement of random noise and its experiment

Abstract

A strictly stationary random process seldom exists in practical random phenomena such as urban noises within a long time interval. Furthermore, noise data within a long time interval has recently been automatically measured in the form of digital level and discrete time. Accordingly, in order to predict and/or estimate the state variables in non-stationary urban noise, it will be necessary to find a skillful method of data analysis by use of digital computer. Concerning the actual road traffic noise with many different types of level distribution forms, at first a new trial of a statistical treatment to estimate the noise level distribution within a long time interval is considered on the basis of any information of the noise level statics (with a high degree of experimental reliability) within a short time interval. The above estimation technique is suitable to the form of digital level and any finite number of data. Finally, the validity of the above estimation theory is experimentally confirmed by applying it to the data of actually measured road traffic noise. The experimental result is in good agreement with the theory.