A new adaptive notch filter with constrained poles and zeros using Steiglitz-McBride method

In this paper we present a new adaptive notch filter (ANF) using the well-known Steiglitz-McBride method (SMM) for an IIR filter with the constrained poles and zeros. The proposed ANF, termed as SMM-ANF, converges to the unbiased solution, has fast convergence speed, and requires less computational complexity than existing recursive maximum likelihood adaptive notch filters (RML-ANF). In the stationary environments, we analyze SMM-ANF convergence properties using the ordinary differential equation (ODE) technique; we derive conditions for the SMM-ANF convergence solution unbiased. Simulations further display that SMM-ANF has better resolution in identifying frequencies of multiple sine waves than RML-ANF. In the nonstationary environments, we also show that SMM-ANF and RML-ANF have approximately identical tracking performance. Simulations are also done to verify the theoretically derived results