Quasi optimal reconstruction algorithm based on the clipping for Gaussian processes

Abstract

A quasi optimal reconstruction algorithm based on the clipping is analyzed for Sampling-Reconstruction Procedure of realizations that compose a Gaussian process. Clipping means that it knows just the zero crossings in the realization. To find out its effectiveness, it is compared with an optimal algorithm, which considers some samples of the realization located at strategic points. Results show that the quasi optimal algorithm does not give a correct reconstruction. Hence, it is necessary to include a new parameter within this methodology to improve the performance, mainly reflected in the reconstruction error. However, the quasi optimal algorithm is just an approximations to the optimal algorithm.