Support vector regression-based robust frequency estimation algorithm by instantaneous phase

There are two important factors which impact the performance of phase-based frequency estimation algorithms remarkably: the approximations of noise phase model and imperfections in phase unwrapping process. Support vector regression (SVR) exits excellent capabilities for learning unknown data and forecasting the future ones, especially under the small sample condition. Therefore, the authors introduce it to predict the variation trend of instantaneous phase and unwrap phases efficiently. Even though with that being the case, errors still exist in phase unwrapping process because of its ambiguous phase characteristic. Furthermore, a SVR-based frequency estimation algorithm is proposed and makes it immune to these error phases by means of setting the SVR´s parameters properly. The results show that, compared with other phased-based algorithms, not only does the proposed one maintain a wide estimation range and quality capabilities at low frequencies, but also improves accuracy at high frequencies and decreases the impact with the initial phase. The proposed algorithm is fit for not only linear phase signal but also polynomial phase signal, under both the Gaussian and non-Gaussian condition.