Using linear prediction to design a function elimination filter to reject sinusoidal interference

A filter structure is proposed for canceling sinusoidal interference with unknown parameters in the presence of a bandwidth noise. This filter algorithm is available over the time-limited observation interval and it does not require any auxiliary reference input. The major part of this paper is devoted to the sensitivity analysis of the proposed algorithm by studying the change in the signal-to-sinusoidal interference ratio with respect to two cases: one being due to a change in additive noise level and the other due to a change in the sinusoidal interference parameters. The experimental results obtained by software filter implementation confirm the canceling of strong sinusoidal interference also in the presence of bandwidth white noise.