Precision and accuracy of the orbital parameters derived from 2D & 1D space observations of visual or astrometric binaries

Recent investigations (e.g. Han et al. 2001) have shown that fitting the Hipparcos observations with an orbital model when the astrometric wobble caused by the companion is below the noise level can have rather unexpected consequences. With new astrometric missions coming out within the next ten years, it is worth investigating the orbit reconstruction capabilities of such instruments at low signal-to-noise ratio. This is especially important because some of them will have no input catalogue, thus meaning that all the orbital parameters will have to be derived from scratch. The puzzling case of almost parabolic orbits is also investigated.