Sometimes I think about not having the classical robotics education in my bag of tricks (educator and Cognitive Psychologist over here, so I’ve spent much more time thinking about humans, rats, philosophy and AI than about Mindstorms in my professional past).
I wanted to learn more in order to better discuss robotics with whip-smart high school students competing in robotics challenges. Most of these students, somewhere along the way, build maze solvers. Why? Well, that’s a little hard to package into a neat, one-sentence answer, but it has a lot to do with Cognition and AI and engineering those qualities into devices.
Here’s our version, with a twist – most robotics is “top down” with a central brain or program controlling sensors and actuators. Cubelets are a parallel and distributed system. So, we set out to make a maze-solver with a plan to use the Bluetooth Cubelet and to program a Cubelets maze solver.
Follow this recipe:
1. First, work at the coolest robotics toy company ever.
2. Build a robot with Cubelets.
2a. Worry that Cubelets don’t have central brains.
3. Decide to reprogram your robot. (This is where the Bluetooth Cubelet comes in)
4. Assemble a crack team to build a maze, define the problem, and iteratively code and test. Spend several hours on this.
5. That night, run this by your best friend who reminds you that often the best way out of a maze is to keep one hand on the wall at all times and just keep following that.
6. Say that to the co-founder of the company.
7. Wait 10 minutes
8. Take this video of the “brainless” robot he built successfully solving our simple maze (without re-programing!). Huzzah!