Active Noise Reduction with Filtered Least-Mean-Square Algorithm Improved by Long Short-Term Memory Models for Radiation Noise of Diesel Engine

This study presents an active noise control (ANC) algorithm using long short-term memory (LSTM) layers as a type of recurrent neural network. The filtered least-mean-square (FxLMS) is widely used ANC algorithm, where the in target area reduced through signal generated from adaptive filter. Artificial intelligence can enhance reduction performance for specific applications. An LSTM artificial network recognizing patterns arbitrarily sequence data. In this study, controller consisting based on deep networks was designed predicting reference signal, which to generate minimize residue. structure and procedure training were determined. Simulations conducted compare convergence time performances with those conventional FxLMS algorithm. source adopted sounds single-cylinder diesel engine, while noises selected single harmonics, superposed impulsive signals engine. characteristics each examined Fourier transform analysis results. simulation results demonstrated that proposed method showed outstanding capabilities narrowband, broadband, environments, without high computational cost complexity relative