How to find a planet
Planets may be big objects on scales that we are used to as humans, but they are really quite small and difficult to detect as they emit no light of their own. This is one reason why 2003 UB313 was only discovered recently. But if you think it is tricky finding planets in our own solar system, try looking for them around other stars. Although many extra solar planets have been discovered in recent years, most of them have been detected by methods such as photometry (a star dims as a planet passes in front of it, requires good alignment of star and planet), spectroscopy (the planet causes the star to wobble slightly which causes a slight doppler shift in the spectrum of the star) and astrometry (directly detecting the wobble of the star caused by the orbiting planet). All of these methods are most sensitive to so-called "hot Jupiters", planets much larger than the Earth which orbit quite close to their parent star.
One way of finding Earth-sized extra solar planets is to use the transit technique from space where the atmosphere causes much less of a problem. The Kepler satellite, due to launch in 2008, is a mission designed to do just that. It will observe many stars in our neighbourhood of the Milky Way watching for the tell-tale dimming of these stars caused by transiting planets. Kepler will be sensitive to Earth-mass planets orbiting at a similar distance from their parent stars as we are from our own Sun.
Another method was described in a talk at a conference in Beijing in June 2005. This is a ground-based technique which makes use of small (1 to 2 metre) telescopes located at Dome C in Antarctica. This method (decribed in the astro-ph paper) makes use of a phenomenon called gravitational microlensing where the light from a distant star is bent, or lensed, by another star which is relatively nearby. If a planet happens to be orbiting the nearer star, then it's gravitationl field distorts the image of the background star even more. The extra distortion of the light from the background star tells us about the presence of the planet even though we cannot see it directly. This extra distortion occurs without the precise alignment of the planet's orbit required for the transit technique.
To see a microlensing event you have to be very lucky. The chance of a nearby star passing directly in front of a distant star is very small, so you have to observe a lot of stars to increase the chance that you will catch one of these events. That's why the reserchers plan to look towards the Galactic centre where there are lots and lots of stars. Current microlensing searches (e.g. OGLE, MOA) observe large fields of stars very regularly to increase their chances of spotting an event.
Why Antarctica? Well, there are several reasons. Firstly, to observe the Galactic bulge you really need to be in the Southern hemisphere as it never rises very high above the horizon from here. Dome C also has very good seeing conditions which are vital for this sort of experiment. During winter you also get uninterupted darkness allowing round the clock observations, vastly increasing your chances of catching a microlensing event.