lab notebooks

robotics test bench

question → prediction → experiment → observation → model update.

the experiment sequence

six preserved experiment records

feedback control, coupled dynamics, jacobians, contact, and leg reachability. the original sequence and its source provenance stay intact.

open the experiment index →

incomplete learning scaffold

ant policy learning

an editable objective-to-behavior scaffold with TODO cells and no saved outputs. this is not a completed experiment and does not receive a public experiment ordinal.

inspect the working notebook →

from isolated experiments to a robot

C-1N is the integration project. its checkpoint explorer preserves earlier standing and walking behavior alongside the accepted STRIDE checkpoint.

inspect C-1N ↗