More slot car hacking

The madness of putting together all of the material and demos for JavaOne is slowly calming down as things come together. The slot car race programming challenge has been sucking up a lot of my time (particularly the software infrastructure). This isn't a challenge that MATLAB is well suited for: it would be if there were accelerometers on the car, but there aren't. It can be used, and I have seen some cool tools to transform MATLAB into RTSJ code. It's also not entirely possible to just precompute a voltage profile and blindly follow it: The physical world is unpredicable in all kind of small ways: tires slip; engines heat up and become less efficient; the track gets polished; there are slight differences between the cars... A winning entry will have to have some degree of adaptive behaviour.
April 30, 2006