One of the star attractions at the recent bring-a-hack prior to our London unconference was [Dan]’s two-wheeled self-balancing robot. As the assorted masses of the Hackaday readership consumed much fine ale and oohed and ahhed over each others work, there it stood on a pub table, defying all attempts to topple it.
In a way a successful self-balancer can look surprisingly unexciting because it achieves the seemingly unimpressive task of just standing there and not doing much except trundling about, but to take such a superficial view belies the significant feat of engineering that gives the self-balancer its party trick. And it’s no mean achievement to create one from fairly basic hardware, so how has he done it?
The 3D-printed frame holds a pair of stepper motors to do the hard work, while a piece of stripboard acts as carrier for boards containing the MPU6050 accelerometer and DRV8825 stepper motor drivers. Meanwhile the brains of the whole show started as an Espruino Pico but has since been moved to an ESP32.
There is a linked GitHub repository with all the code, and if our description of seeing it in a London pub isn’t good enough for you then you can see it in action in the video below.
Continue reading “A Robot That Can Still Keep Its Balance After A Night In The Pub”
Since then it has been completely redesigned and now has a super fat kite-surfer wheel, a front crash skid with damper, and a variable geometry which allows it to steer properly despite just having one wheel. It does this by allowing the rider to shift their position relative to the wheel, instead of the seat always being rigidly locked directly above the axle.
