Signal processing system using a neural network for laser Doppler velocimetry

A signal processing system using a neural network for the laser Doppler velocimetry is described. The proposed system is based on spectrum analysis using a fast Fourier transform and adaptive prediction using a neural network. The neural network learns fluctuation of the velocity using the backpropagation algorithm and predicts the velocity related to the Doppler frequency adaptively. Actual measurement on a rotating acrylic disk was carried out. An increase in the accuracy of the velocity measurement was obtained.<
>