Instantaneous frequency estimation: confidence bounds using the bootstrap

This paper uses a nonparametric method for estimating confidence intervals for the instantaneous frequency. The method, known as the bootstrap, is computer-intensive but requires very little in the way of modelling and assumption. The confidence intervals are based on the least-squares instantaneous frequency estimator. Simulation results demonstrate the potential of the method in situations where theoretical analysis cannot at all, or only with considerable effort, provide the results.<>