A signal processing method is proposed whereby the frequency of a noisy sinusoidal signal can be estimated optimally. The criterion for optimum performance is the minimum observation time for a given error or the minimum error for a given observation time when the same signal-to-noise condition prevails. Three methods of analyzing the instance of the transition of the signal through a selected level in a given direction can be used to determine the frequency. These methods have been simulated and the results give a comparison of the measurement errors for the various methods, when using different signal-to-noise ratios and observation times. The third method shows considerably better results than the first two. © 1990 IEEE
