After 5,000 iterations, Microduck robot still fails to walk reliably

tristanbob · x · 2026-08-30

The author shares results from 5,000 RL training iterations on the Microduck robot using default settings: the model failed to learn reliable forward movement, instead discovering that head movements (looking sideways or down) help initiate motion. The author notes the difficulty of RL and plans to experiment with new parameters and rewards next.

Related event: RL-Trained Virtual Duck Fails to Walk Reliably After 5,000 Iterations(6 posts)→

Original post →

More from Embodied

Embodied channel →