ResearchBenchmarksRL Motion Imitation on Unitree H1-2 One-foot BalanceFollow360Final RewardDDR249.448278.149306.85335.551May 22, 2026Evaluation ResultsMethodMethodLinksFinal RewardSteps to 90% RewardDDRReference Type=Direct...Reference Type=Direct Dynamic Retargeting2026.0536076,300IDRReference Type=Immedia...Reference Type=Immediate Physics Correction2026.05301.657,300GRReference Type=Kinemat...Reference Type=Kinematic Retargeting2026.05253.7219,900