Skip to content

Project: SO-101 Arm

Manipulation driven by learned policies: teleoperate the SO-101 arm, collect demonstration datasets, train imitation-learning policies, and run them on real hardware.

Milestones

  1. Teleoperation and dataset recording working end to end
  2. A trained policy completes a pick task autonomously
  3. Policy generalizes to object positions not in the training data

Iulia's Hardware Picks

Some links below are affiliate links. As an Amazon Associate I earn from qualifying purchases. I may also earn a small commission from other links, at no extra cost to you, and I only list gear I actually use and recommend.

Watch

See in tree →

Next in the tree