Basically you have to start simple, and use something they do naturally (in this case, run around to search for food). Except in this case we put them on top of a ball treadmill, so that they move around in a virtual environment instead of a real one. Here’s a cool video from a group that uses almost the same technique as we do:
They very quickly get the idea though that ‘when I run through the stripey thing I find a drop of soy milk’. And from that point on you can start making things slowly more complicated (make the targets move around, add ‘bad’ targets etc.)
Comments