Frequency estimation from short pulses of sinusoidal signals

This paper treats the problem of estimating the frequencies of sinusoidal components from short, incoherent, noisy pulses. We formulate the problem in a stochastic framework and construct a generalized Yule-Walker solution by minimizing a cost function. This approach gives an improved estimate of the covariance matrix and is computationally efficient. Noise is mitigated by incorporating the principal component analysis algorithm. A computer simulation applying this algorithm to FSK data demonstrates that the proposed algorithm converges more rapidly and with more accuracy than a competing method.