HULK: Learning Whole-Body Forceful Loco-Manipulation for Humanoids

Key points

• Humanoid loco-manipulation of large, heavy objects demands forceful interaction across the entire body.1
• However, such payloads shift a humanoid's center of mass and impose sustained loads across the upper body, challenging balance and command tracking.1
• Using model predictive control (MPC) to guide reinforcement learning with predictions of the loaded dynamics, we train two teachers: one tracks arm motions under wrist forces, and the other locomotes while holding large objects against the body.1
• A capture-point control barrier function augments the wrist-force teacher during training to improve balance under load.1

Sources (1)

• [1]HULK: Learning Whole-Body Forceful Loco-Manipulation for Humanoids
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 06:33 PM

  We present HULK, a whole-body control framework for forceful loco-manipulation.

  Humanoid loco-manipulation of large, heavy objects demands forceful interaction across the entire body.

Extractive summary: sentences quoted from the sources.