Study On Sinusoidal Frequency Estimation Based On Maximum Likelihood Estimator

Sinusoidal frequency estimation is the procedure which can estimate the frequency of sinusoidal waves in additive white Gaussian noise through transform and calculation for signal's samples. Not only in theory, but also in practice, the sinusoidal frequency estimation has essential study merit.The maximum likelihood estimator (MLE) approach is an important one in frequency estimation domain. This approach has the merit such as exhibiting best performance and requiring minimum threshold of signal noise ratio in calculation. However, as the MLE requires obtaining the parameters of the high order non-linear likelihood function when it reaches the global maximum, this makes the MLE approach has very huge calculation burden.In this paper, when the global optimal theory and Monte - Carlo technique are applied, I transformed the maximum likelihood frequency estimation problem of sinusoidal waves from multidimensional grid search to a closed expression. This makes it easy to estimate the frequency through optimal calculation.In my paper, I put emphasis on MLE frequency estimation approach and it's implementation procedure. Over and above, I put forth a new solution for frequency realization algorithm after mending it. I also give a new way to resolve the estimation problem of amount of sinusoidal waves that is an unresolved problem nowadays.At the end of my paper, I tested my new algorithm and solution by simulation and proved them effectiveness and correctness.