1998 Volume 118 Issue 1 Pages 29-36
The adaptive frequency estimation has many engineering applications, and a lot of techniques have been worked out. In this paper, we propose a new computationally efficient adaptive frequency estimation algorithm which is based on the blocked variable sampling and the synchronous addition / subtraction processing (SAS). The SAS requires additions and subtractions only, which makes the new frequency estimation technique very attractive in terms of computation load. Extensive simulations are presented to reveal the performance of the proposed technique which is expected to have potential applications where the sinusoidal frequencies vary slowly around their nominal values and the estimation algorithm with small complexity is highly desired. The proposed algorithm is also compared with the DFT (FFT) and the conventional gradient-based IIR adaptive notch filter methods in terms of both estimation accuracy and computational complexity.