Resuming this thread after a long break: I successfully trained an ACT policy to pick the duckie and place it on top of the stand: it works!

On Day 3 I had the opportunity of beta-testing MakeDevice, an awesome tool in development at the University of Lancaster that will blow away the barriers and enable anyone to build electronic devices #pro2summerschool2025

Next, design an enclosure made of laser-cut acrylic using Onshape, and build it #pro2summerschool2025

The sloppy soldering is entirely my fault, but it works!! #pro2summerschool2025

Adding volume to the icing on Day#4

Enjoying a fun day at the #FirstLegoLeague in Seville with the little one.

150 kids aged 7 to 16 showing off their awesome projects and battling it out in the robotics comp - what an experience!

Teams came from Andalusia, Extremadura, and the Portuguese Algarve.

This is where the future starts!

Day 14 (part two): Last but not least, the assembly of the handle. I love the grip in it. Looking forward to give it a try for teleop.

Day 14: assembly of the leader arm is now complete! (And video edition much more efficient). Part 1 of 2.

A bit more visual

I’m looking at using a Nintendo Switch to control my robots. Joy-Cons promise to be fun, and I had a first breakthrough today. A short video, and details in the repo: github.com/mhered/ninte...

Day 12 (cont’d): part 2 of the video. Editing 2h of raw video down to 2’ took me days, but I am happy with the result.

Day 12: the assembly of the leader arm begins. Splitting the video in two due to bsky’s 1’ limit

Day 11: playing with URDF Visualizer VSCode extension

Day 9: setting up the servos of the leader arm with the waveshare board and the script provided by lerobot - a minimal video tutorial

Day 8: Back to 3D printing. Note how the choice of electric green for the leader arm exudes understated,
sober elegance 🤪

Day 7: this thread I came across searching in Discord reinforces my suspicion that to get rid of the jittering I may need to play with PID parameters: x.com/ilialarchenk...