ETH team teaches a 5-tonne skid-steer loader to balance and drive on two wheels autonomously
lukas_m_ziegler · x · 2026-11-18
A team from ETH Zurich's Robotic Systems Lab (leggedrobotics) trained a policy that autonomously performs a stunt only a handful of elite human operators can do: tipping a skid-steer loader onto two wheels and driving balanced.
The machine is an operational 5-tonne John Deere 332G. The policy swings it up onto two wheels, balances while driving over rough ground, traces set trajectories, spins, swings the bucket, then sets it back down on four wheels — all on its own.
Two things make it hard: hydraulics introduce flexible elements and nonlinear delays that are nasty to model and control, and a 5-tonne machine at its limits is far too dangerous to learn by trial and error. The post is truncated before full training details.
More from Embodied
- Musk predicts Tesla Optimus will wash, dry and put away dishes in 12-18 months — Polymarket · 2026-11-19
- Nimo Robotics launches N1 glove and D1 hand: highest-DoF commercial UMI system for industrial robot skills — Scobleizer · 2026-11-19
- Robot Model Dream-1 Learns to Drive a Real Tesla From 1.5h of Demos — scott_e_reed · 2026-11-19
- Stack-chan desktop robot gets a screwless multi-leg armor mod, files open-sourced — alexcovo_eth · 2026-11-19
- Inceptio tops 1.2B km of driverless-truck operations, makes the case for freight-native physical AI — 量子位 · 2026-11-19
- Antigravity Robotics plugin lets anyone build MuJoCo sim tasks with prompts — DynamicWebPaige · 2026-11-19