Abstract
Active noise control (ANC) is achieved by introducing a canceling “anti-noise” wave through a control source. The anti-noise of equal amplitude and opposite phase is generated and combined with the undesired noise based on the principle of superposition, thus resulting in cancellation of both noises. Although the idea of ANC was first proposed in 1936, the practical applications were successfully implemented in recent decade due to the available of low cost and efficient digital signal processor (DSP). ANC has been widely applied to transportations, manufacturing, and consumer products. This thesis will focus on the practical aspects of ANC system in utilizing modified filtered-x NLMS(MFxNLMS) algorithm for real-life applications. The normalized step size can be updated based on the power of reference signal, the convergent speed of this proposed method will faster than traditional modified filtered-x LMS(MFxLMS) scheme. Experiments show that MFxNLMS not only cancels narrow-band and broadband signals quite effectively, but also selectively cancels signals with pre-specified frequency or pre-specified band.