I'm Stéphane Caron, a robotics researcher at Inria Paris working on visuomotor control.

During my work, I have developed motion control software for HRP-4 humanoids (open source), ANYmal quadrupeds, and now Upkie wheeled bipeds (open source). You can read the algorithms in the corresponding publications, and check out the code on GitHub. I also toot and write technical notes on this website.

Robot locomotion

One way to develop robot behaviors in the lab is to aim for repeatable demos, but at some point this goal steers us away from generalizing to the real world. For instance, the LIPM walking controller I worked on for the HRP-4 humanoid could repeatedly achieve a relatively complex task like stair climbing, but it wouldn't know how to react to unscripted scenarios like banging its head on a ceiling. That was due in part to all its behaviors being implemented by control systems, each coming with its own set of parameters to tune, and the complexity of "keeping it all working together" increasing significantly with each new behavior. There are several ways we can explore to cross this complexity ceiling...

To help with experimenting faster (and alleviating the nervousness of breaking big expensive robots), I made Upkie, a wheeled biped that is fully open source and can be built following step-by-step instructions. Upkie is a fast-prototyping partner. Most of its code runs in Python on a Raspberry Pi, which is enough for tasks like balancing and locomotion. Upkie's software works out of the box with PID control, model predictive control and reinforcement learning.

Open access

I support open access and was fortunate to be able to release most of my research as pre-prints and software that I can keep maintaining. On top of open access, I see three ways in which we could work better: overlay journals, maintaining post-prints, and raising our standards in terms of code distribution.