Nonuniform sampling and polynomial transformation method

Nonuniform sampling may be preferred over uniform sampling for various reasons. The polynomial transformation method which converts the nonuniformly sampled data into signal values at uniform spacing for further processing has some desirable properties. The study shows that the transformation method may be advantageous even for the case where the noise-corrupted signal is sampled at uniform intervals. It is demonstrated that, by employing the transformation method, the accuracy of the solution of the classical problem of parameter estimation of complex exponential signals can be substantially improved.