Engineers at Johns Hopkins University have created a set of robotic legs that can be controlled using brain signals.  The legs are controlled through a electronic cap so the technology is non-invasive.

So far, the device is able to recognize when a person wants to walk or stop with 95% accuracy.  All the user has to do is imagine walking or stopping and the legs behave accordingly.  Using EEG signals, paralyzed people throughout the entire world will one day be able to go about their day again with just the effort of a thought.

Research has also made advancements using EEG signals to control a computer cursor and/or drive a wheelchair.

The legs still need to have speed and direction factored into their movement, but this is a few steps in the right direction.