Novel statistical offline nonuniform sampling of time-limited signal

In this paper, a novel statistical approach for nonuniform sampling of a time-limited signal is proposed. Based on dynamic range, the signal is divided uniformly in equal intervals which constitute coarse partition. Then, total number of samples is nonuniformly distributed among the intervals of coarse partition leading to nonuniform sampling. The problem for nonuniform allocation of samples within coarse partition has been formulated and solutions based on integer linear programming have been proposed. Using intelligent allocation of samples in nonuniform manner, fewer samples are required to represent the signal retaining maximum information of the original signal. Approximation error between uniform sampling and proposed nonuniform sampling is compared. Original time-limited signal is reconstructed using B-Spline interpolation function.