Hacker News new | ask | show | jobs
by Matti 1804 days ago
Classical control theory is often sufficient for dealing with changing loads, terrain and wear and tear. Rejecting unknown disturbances is what a control system does. No training is required. Perhaps somewhat less than impressive due to the choice of robotics platform, but:

Achieving natural behavior in a robot using neurally inspired hierarchical control Joseph W. Barter, Henry H. Yin Paper: https://www.biorxiv.org/content/10.1101/2021.01.22.427862v1

Movie 8: https://www.biorxiv.org/content/biorxiv/early/2021/01/25/202...

Movie 9: https://www.biorxiv.org/content/biorxiv/early/2021/01/25/202...

1 comments

Yeah, a lot of this tends to be "Let's try CNN on ... ", and it ends up doing about as well. It's bound to be a general purpose tool that's easy to use and doesn't require a classical approach.

Like RRTs instead of visibility graphs for regular-old polygonal environments. The RRT is just easier vs hand-tuning obstacle boundaries, etc.

I have some experience on this, as we are building a system like this with similar constraints, and there's really no "need" for anything other than classical control.