Title :
Approximation error for the estimation of a continuous unknown deterministic signal in noise from discrete time samples
Author_Institution :
New Mexico State University, Las Cruces, New Mexico
Abstract :
The approximation of a continuous unknown deterministic function from samples of the function imbedded in noise is considered. Using the integral mean squared error criterion, the total approximation error composed of a deterministic and random part is found for the step function, straight line, and cubic spline approximations. It is shown for a gaussian function that if the noise variances are equal and larger than a specified threshold, then the straight line approximation is superior to the spline approximation while both spline and straight line approximations were always better than the step function approximation.
Keywords :
Approximation error; Computer errors; Estimation error; Function approximation; Gaussian noise; Interpolation; Joining processes; Missiles; Spline; State estimation;
Conference_Titel :
Decision and Control including the 13th Symposium on Adaptive Processes, 1974 IEEE Conference on
Conference_Location :
Phoenix, AZ, USA
DOI :
10.1109/CDC.1974.270410
