Mixed-error approach for multi-channel active noise control of open windows

This paper presents an attempt to reduce the computational complexity, while achieving acceptable level of noise reduction in a multi-channel active noise control (MCANC) system for mitigating noise passing through open windows. The reference signals are sensed at the back of the secondary loudspeakers. The existing approach to reduce the computational complexity has already used only one reference signal in each feedforward channel. As the number of error microphones is also important to ensure global noise reduction of the MCANC system for open windows, the main idea of this paper is to preprocess the error signals so that the number of inputs to the controller is further reduced. When the preprocessing is a simple summation, it is called the mixed-error approach. The number of inputs related to the error signals is eventually reduced to 1 after the mixed-error approach is applied. Simulation results demonstrate that the optimum control filters derived under general conditions can lead to similar global noise reduction to those derived based on the mixed-error approach. Experimental results confirm that in a 4-channel MCANC system applied to a 0.2 m by 0.2 m open window, the performance of global noise reduction is not compromised by the mixed-error approach.