摘  要：Detection of Earth-Twins is the one of the most exciting hot topics
in modern astronomy and in public domain. So many observatories, so many space missions,
and so many different techniques, including RV, photometry, imaging, occultation,
interferometers, adoptive optics etc., have been dedicated to detection of exoplanets.
So far hundreds of exoplantes have been discovered and a few of them are claimed
to be super-earth planets. The fact is, however, that most of them are candidates of
extra-solar planets only, and need further verification simply because of that their
fundamental physical quantity, mass, is unknown. The precision RV technique may reach
10 cm/s for instrument errors. But the astrophysical noises from star spots, pulsation
and intrinsic RV variations prevent the possible detection of nearby Earth-like planets.
In addition, some missions may provide statistics of Earth-like planets and cannot
detect Earth-like planets for nearby stars which are extremely important for future
imaging and spectroscopic investigations. Space astrometry, however, is the only
technique, which can have precision of 0.1 μas to detect Earth-Twins, and can provide
orbits and ephemeris of Earth-Twins for nearby solar-type of stars.

I will present latest results of detection of exoplanets and demonstrate error estimates
of different techniques. In particular, I am going to show interferometric precision
and accuracy for wide-angle and narrow-angle astrometry. I will compare different space
astrometry missions and discuss challenges and status of space astrometry for detection
of Earth-Twins.