EFFICIENT IMPLEMENTATION OF MULTIHARMONIC LEAST-SQUARES FITTING ALGORITHMS

Abstract

Multiharmonic least-squares fitting algorithms have been applied in many fields of instrumentation and measurement. However, as the number of harmonics increases, the size of the matrices needed to describe the relation between the sampled signal and the multiharmonic model increases substantially. In this paper an efficient memory-wise implementation of the algorithm is proposed, and numerical simulations are used to assess its performance. Measurement results are also presented.