Active Noise Control Using Intrinsic Mode Function Technique

Active noise control accuracy depends on how much destructive interference exists between the primary noise and the noise (anti noise) generated by secondary source. In this paper firstly multilayer perceptron (MLP) neural network is designed and trained with extracted features of noise signals for classification and then Empirical Mode Decomposition (EMD) is used for active noise control. The noise signals of m109 tank and F16 cockpit are selected from SPIB database. The results of simulation show that the EMD technique is capable of suppressing the non linear and non stationary noise signals. The EMD technique has performed well in noise attenuation.