A stable adaptive notch filter with optimal tracking properties

In this paper, an analysis of the properties of adaptive notch filter (ANF) applied to direct frequencies estimation is presented based on the results of G. Li (1997) and M.V. Dragosevic and S.S Stankovic (1995). The ANF presented in G. Li (1997) is purely parameterized by notching frequencies of the filter. This characteristic makes it not only much simpler but more robust compared with traditional algorithms. ANF presented in M.V. Dragosevic and S.S Stankovic (1995) that includes adaptation of both pole contraction and forgetting factors results on superior tracking properties. On the basis of such results, we present a stable ANF with improved tracking characteristics. In experimental part, we use mentioned method to track stationary single sinusoid, frequency-hopping single sinusoid, in additive white noise. Simulation results confirm the theoretical conclusions and show that they are highly efficient in practice.