Estimating bi-directional Doppler flow parameters using time-frequency distributions

Estimating the mean speed of a bi-directional flow is one of the main objectives in Doppler flow measurement. It is known that the mean speed in the unidirectional flow is proportional to the pseudo instantaneous mean frequency (PIMF) of the Doppler signal acquired through a flow measurement process. Two different approaches are presented in this paper to compute the PIMF of a bi-directional flow signal, using time-frequency distributions (Cohen's class). The first of them consists on extracting the forward and the backward flow components of the Doppler signal considering them as separated unidirectional flows. This involves the use of their analytic signals and the calculation of the conventional PIMF. The second method does not separate the flow components. Instead, a generalized PIMF is used to analyze the quadrature signal acquired adequately. Results show the performance of both approaches under certain conditions, in particular using the Wigner Ville TFD and the flow separations Phasing Filter Technique.