Active noise reduction method and system based on deep recurrent neural network, and storage medium

Abstract

The invention discloses an active noise reduction method and system based on a deep recurrent neural network, and a storage medium. A reference sensor transmits a detected noise signal to a digital signal processing unit; the digital signal processing unit converts the noise signal into a reference signal and transmits the reference signal to a controller unit; the controller unit constructs a recurrent neural network model with a hidden layer containing a gating function by using a long-short-term memory unit structure; an output signal of the controller unit is transmitted to the actuator unit after being processed by the D / A conversion module, and the actuator unit produces sound to generate a secondary sound source; a noise source generates a primary signal at the error sensor after being propagated through the primary path, and the generated secondary sound source generates a secondary signal at the error sensor after being propagated through the secondary path; and an error sensor receives the residual signal and feeds back the residual signal to the controller unit, continuously updates the weight coefficient of the DRNN structure in the controller unit according to the target function, and realizes noise reduction by minimizing the target function.

Description

technical field [0001] The invention belongs to the technical field of noise reduction, and in particular relates to a nonlinear active noise reduction method, storage medium and system based on a deep cyclic neural network. Background technique [0002] Active Noise Control (ANC) has become a viable technology and has been successfully used in a variety of applications. Most of the traditional ANC systems using linear filters as controllers use adaptive filtering algorithms. The filters automatically adjust their own weight coefficients through adaptive algorithms to minimize the objective function of the system. For linear active noise control The system has good stability and reliability. However, since the linear ANC system does not consider the actual ANC system components and the nonlinearity contained in the acoustic field, the adaptive algorithm will hardly achieve good control effect. [0003] Neural network is a nonlinear control technology developed in recent ye...

AI Technical Summary

Problems solved by technology

However, the actual active noise reduction system is a discrete time system. The analog signal collected by the sensor is preprocessed and converted into a digital signal as the system input. The input and output are both time series signals. The traditional neural network structure cannot learn the noise signal well. time correlation

Images