Neural networks applied to continuous range amplitude measurements of small sinusoidal vibrations [using Michelson interferometer]

A novel method for computing the amplitude of small sinusoidal vibrations measured with the Michelson interferometer is proposed. The method is a particular case of nonlinear adaptive inverse modeling, using neural networks. The feasibility of the method is proved by simulation, and the noise behavior is examined. The results are of practical interest: they show that the method is applicable to practical, noisy signals and the precision can be very good.