Noise Parameter Estimation in the Presence of Random Timing Errors

Algorithms for estimating the noise parameters of a signal sampled with random timing errors and embedded in additive noise are proposed. Both the timing errors and the additive noise are assumed to be Gaussian and independent and identically distributed. Computationally efficient estimators, derived by maximising an approximation to the likelihood, are proposed. Comparisons with the Cramer-Rao bound demonstrate the performance of the proposed algorithms for a range of parameter values