Hi team, great work on AMO. We're deploying a teleoperation pipeline on our G1 with PICO controllers.
The AMO policy generally works, but we often see stability issues — the robot doesn't fall, but frequently takes large involuntary steps trying to recover balance and we need to hold it.
Did you experience similar stability issues, or was the policy consistently stable?
Any recommendations for improving stability (e.g., specific height/pitch range limits or operating conditions that helped)?
Hi team, great work on AMO. We're deploying a teleoperation pipeline on our G1 with PICO controllers.
The AMO policy generally works, but we often see stability issues — the robot doesn't fall, but frequently takes large involuntary steps trying to recover balance and we need to hold it.
Did you experience similar stability issues, or was the policy consistently stable?
Any recommendations for improving stability (e.g., specific height/pitch range limits or operating conditions that helped)?