A noise cancellation method in sound and electromagnetic environment of power state variables

Abstract

The observed phenomena in actual sound and electromagnetic environment are inevitably contaminated by the background noise of arbitrary distribution type. Therefore, in order to evaluate sound and electromagnetic environment, it is necessary to establish some estimation methods to remove the undesirable effects of the background noise. In this paper, we propose a noise cancellation method for estimating the power state variables of a specific signal with the existence of background noise of non-Gaussian distribution from two viewpoints of static and dynamic signal processing. By applying the well-known least mean squared method for the moment statistics with several orders, a practical method for estimating the specific signal is derived. The effectiveness of the proposed theoretical method is experimentally confirmed by applying it to estimation problems in actual sound and magnetic field environment.