Multiple reference active noise control - The attainable suppression

Active noise control is a well-known technology for the attenuation of disturbing sound or vibration. Feedforward control is efficiently employed, especially if stochastic disturbances are to be suppressed. If the unwanted noise is generated by multiples sources, multiple reference sensors are also used. The behavior of multiple reference feedforward systems has been investigated in the last decades, with special emphasis on the speed of convergence. It has been proven that reference preprocessing can significantly speed up such systems. However, less attention has been paid to the attainable suppression. This paper provides the analysis of multiple reference systems with respect to the steady state, and an upper bound for suppression is also expressed. Besides, the single reference case, and the influence of preprocessing is discussed. The theoretical results are supported by simulation and experimental examples.