Paper: arxiv.org/abs/2511.05033
Documentation: gatech-epic-power.github.io/epically-pow...
GitHub: github.com/gatech-epic-...

Please feel free to reach out, drop a note in our GitHub Discussions page, etc. We're really looking forward to getting feedback, iterating on this, and seeing how people use it - both in wearable and broader areas of robotics!

Almost all (if not all) devices in our groups at Georgia Tech are running on EP, which has helped streamline development and keep things consistent across teams. It cleans up mid- and high-level controller implementation, while still giving users behind-the-scenes access if they want/need it.

Why use it?
- Seamlessly interface with a range of actuators, sensors, and single-board computers
- Quickly implement controllers and visualize data in real time
- Customize and scale robotic systems without rebuilding core software
- Swap out components without rewriting entire controllers

We've created this so the code interfaces with all devices in a similar stylistic way, regardless of the varied communication protocols we're handling in the background.

We've covered actuators from three of the big manufacturers, RasPis and Jetsons, various IMU types, and different power options.

Highlights:
- Works with many actuators, single-board computers, and sensors without locking users into pre-defined hardware setups
- Easy-to-use Python interface
- Assembly tutorials and example controller code
- A part selection guide to help find the right components for specific applications

After significant development by our team - and testing from our lab, as well as colleagues at Northeastern, Carnegie Mellon, and Washington - it’s now a robust, modular, open-source platform that makes it faster and easier to go from idea to working robotic system.

This started as an in-lab effort to streamline the essential “plumbing” of wearable robots - low-level communication parsing, clocking, data acquisition, visualization, etc. - tasks that were taking us weeks or months to implement and troubleshoot, and often being redone by new folks in the lab.

@gatechengineers.bsky.social @gregory-sawicki.bsky.social
@amsocbiomech.bsky.social
@blackinengineering.bsky.social @blackinx.bsky.social (2/7)

I'm a current Harvard graduate student and I found out today that I had my NSF GRFP terminated without notification. I was awarded this individual research fellowship before even choosing Harvard as my graduate school

The main benefit of these is that we can study ground stiffness perturbations on a fully instrumented treadmill, as well as conduct experiments overground. I'm hoping to share results from a study investigating adaptation to a stiffness perturbation applied by these shoes very soon!