Coherence beamforming applied to velocity estimation and partially coherent signals

Time-delay and phase-shift estimators are often used to track speckle signals in order to estimate parameters such as strain, displacement, and velocity. These estimators rely on the beamformed RF signal to accurately estimate time and phase shifts. Partially coherent targets such as speckle, however, introduce variance and bias errors into the estimation because the delay-and-sum beamformer does not adapt to the spatial variation of the sampled waveform. In addition, the presence of acoustic and thermal noise can exacerbate this problem. We introduce a method that modifies the information used by the beamformer when it is applied to velocity and displacement estimation. Using simulations, velocity was estimated between RF channel signals, rather than beamformed RF signals, using Kasai´s and Loupas´ autocorrelation methods. Velocity and displacement was estimated from those channel signal pairs that exhibited high correlation. By eliminating incoherent information in Kasai´s and Loupas´ velocity estimators, jitter was reduced in high noise environments. For example, at 0 dB noise relative to the blood in the channel signals, Limiting the signal to high coherence signals (e.g. 10 spatial lags) reduced the jitter in the Kasai velocity estimator from approximately 6.0 mm/s to 2.5 mm/s. In general, jitter increased with increasingly incoherent information added into the beamformed signal and low noise yielded smaller jitter.