Several years ago I attended CES in Las Vegas. I was working on a medical audio project, but did my best to take in as much of the show as possible. In one of the far-flung halls I came across rows and rows of companies selling robotic toys, robot kits, and home robots. I was impressed by the variety of products that were presented as “consumer robotics”.

It was also interesting to me how few of these products included development tools or any computational ability beyond a simple microcontroller connected to basic sensors and motor control. It seemed to me like there was a gap between mass produced robotic toys and heavy duty research robots. Then I wandered over to iRobot’s booth. I had always been skeptical of their products. How could a little battery powered lightweight device be an effective vacuum cleaner, let alone a well developed robot? But when I saw it in action, I was really impressed by what the Roomba represented. Here was a consumer robot that, while not exactly perfect, did an adequate job for a small, automated vacuum cleaner. Given the reasonable pricing, I consider this a first in robotics. I lingered around the iRobot booth long enough to come across another cool product called the iRobot Create.

This looked like a robust, simple, and affordable mobile robot base, with a payload bay for my projects. Awesome! It didn’t take too long for me to get my hands on a Roomba and a Create. The Roomba was immediately work cleaning all sorts of engineering detritus from my office floor, but the Create gathered dust for a long time while I thought of the right project and the right computer to control it with.

As I sat thinking about ideas for the second SPARK project, I realized that the SPARK compatible hardware would make a great robot brain, giving me capabilities that take the Create well beyond just being a glorified RC car. I began thinking about computer vision, speech recognition and generation, and high speed data computation, all coupled to a power efficient and physically compact computing platform. I had to restrain myself a little though. Even though I’ve read about these Windows CE libraries, I’ve never used them. Vision and speech recognition are very challenging to code, even with great libraries. So rather than define my project so completely from the beginning, I am more thoroughly exploring the “getting started” process. As I successfully implement each feature, I’ll move on to more complex functionality.