An Adaptive Noise Cancelling Method Based on EMD for Acoustic in Battlefield

An adaptive noise cancelling method based on EMD (empirical mode decomposition) for acoustic in battlefield is proposed, which can subtract noise from acoustic in battlefield and enhance the identification rate of acoustic target. Selecting four microphones for synchronous signal sampling, signals can be decomposed with EMD into a finite and often small number of IMF (intrinsic mode function). The sum of several IMF components is taken as reference input of adaptive noise cancelling based on the characteristics of EMD. Lots of experiments show that EMD provides an effective way for selecting reference input. For acoustic in battlefield, taking the sum of three IMF components as reference input of adaptive noise cancelling can effectively subtract noises from primary input containing both signal and noise in battlefield. Further experiments show that acoustic in battlefield after adaptive noise cancelling with the method proposed in this paper is more stationary, and can be used for acoustic target classification and identification.