A spline smoothing approach to transient signal reconstruction

The author considers the problem of transient signal reconstruction in a nonparametric regression setting, where L-spline functions are shown to be optimal estimators for a restricted class of transient signals. On this basis, a Monte Carlo simulation is presented to show the performance characteristics for a special case of L-spline functions in the signal reconstruction for several typical transient signals modeled on a white Gaussian noise background. It is shown that one can recover such transient signals, which are contaminated with noise, to a large degree of accuracy