Unitree Go-2

The 'Unitree Go2' is a quadruped robot used to evaluate the fine-tuning of reinforcement learning policies for dynamic locomotion tasks, specifically in jump and trot scenarios.

Papers using Unitree Go-2 (12)

Slowrl: Safe Low-rank Adaptation Reinforcement Learning For Locomotion2026Towards Embodiment Scaling Laws In Robot Locomotion2025RANDPOL: Parameter-efficient End-to-end Quadruped Locomotion Via Randomized Policy Learning2025Mcarl:morphology-control-aware Reinforcement Learning For Generalizable Quadrupedal Locomotion2025MS-PPO: Morphological-symmetry-equivariant Policy For Legged Robot Locomotion2025Towards Embodiment Scaling Laws in Robot Locomotion2025McARL:Morphology-Control-Aware Reinforcement Learning for Generalizable Quadrupedal Locomotion2025RANDPOL: Parameter-Efficient End-to-End Quadruped Locomotion via Randomized Policy Learning2025PGTT: Phase-Guided Terrain Traversal for Perceptive Legged Locomotion2025MS-PPO: Morphological-Symmetry-Equivariant Policy for Legged Robot Locomotion2025UEREBot: Learning Safe Quadrupedal Locomotion under Unstructured Environments and High-Speed Dynamic Obstacles2026SLowRL: Safe Low-Rank Adaptation Reinforcement Learning for Locomotion2026