Essentially, using interferometry in a different way to get binary star orbits where the techniques leave a gap in separations that is difficult to observer.

i.e. a speckle interferometry camera on a 4-m telescope can astrometrically resolve a binary into its components down to around 30-35 milliarcseconds and traditional interferometry really depends on the baseline and measuring visibility up to about 10mas. We have a way to use multiple baselines to observe stars from anywhere around 10mas to over 100 using the 1-D measurements and getting a fringe packet for each component to get the true position.