The so-called EmDrive should not work, according to established principles of physics. Yet lab tests seem to stubbornly and inconveniently show something is going on. The EmDrive was first demonstrated by British engineer Roger Shawyer in 2003, and has been argued about ever since.
Finally someone figured the best way to prove once and for all if the darned thing works was to put one in space and try it out. It's still a big deal putting something into space, so they must figure it has some chance of actually working.
If it does indeed work in space, the EmDrive should be very useful in propelling probes to other planets at high speeds.