Reinforcement Learning for a Self-Balancing Motorcycle

RE: Reinforcement Learning for a Self-Balancing Motorcycle

madhugovindarajan — Fri, 05 Oct 2018 18:54:19 GMT

For those who are interested to learn more about the kit, here is a recent webinar that my colleague Dan and I gave - https://www.mathworks.com/videos/classroom-ready-projects-with-the-new-matlab-and-simulink-based-arduino-engineering-kit-1537279713545.html

RE: Reinforcement Learning for a Self-Balancing Motorcycle

jw0752 — Wed, 05 Sep 2018 15:56:09 GMT

Hi Madhu,

This was a great project. Initially I thought that you were trying to use the common inertial properties of the flywheel to balance the bike. Of course the position of the wheel was not in the correct plane for this. The idea to add or subtract momentum from the wheel as the source of the balancing force was very cool and not something that I had considered before. I thought your presentation was excellent.

John

RE: Reinforcement Learning for a Self-Balancing Motorcycle

Fred27 — Tue, 04 Sep 2018 11:29:06 GMT

An interesting post and a nice kit to experiment with. Of course, a real motorcycle would be far more complicated as often you want it to lean (even relative to what "feels" like down), and to be able to change angle quickly as you corner.

RE: Reinforcement Learning for a Self-Balancing Motorcycle

genebren — Mon, 06 Aug 2018 19:16:50 GMT

Very interesting blog. There is some great information here, that I will have to come back and re-read, once I have thought about this a little more.

Thanks!

Gene

RE: Reinforcement Learning for a Self-Balancing Motorcycle

three-phase — Mon, 06 Aug 2018 18:39:12 GMT

Interesting blog, thanks for posting.

Kind regards