Event
Authors propose ThinkLocoManip legged mobile manipulation framework
1 reports · 1 sources
Nuo Chen and other authors propose ThinkLocoManip, a whole-body control agent framework for legged mobile manipulation, which adopts a paradigm of alternating reasoning and execution to handle long-horizon tasks in large-scale environments. The framework orchestrates perception, navigation, manipulation, and locomotion as composable tools, and designs cross-domain coordination skills and a keyframe thinking mechanism to support error recovery. The authors report that in real-robot experiments the framework achieved a success rate of 77%, more than double the 37.1% of the strongest plan-then-act baseline.
Report timeline
- ThinkLocoManip: A Whole-Body Control Agent Framework for Legged Loco-Manipulation
The authors propose ThinkLocoManip, a whole-body control agent framework for legged loco-manipulation that uses an interleaved reasoning-and-acting paradigm to handle long-horizon tasks in large-scale environments. The framework orchestrates perception, navigation, manipulation, and locomotion as composable tools, and it designs cross-domain coordination skills and a keyframe thinking mechanism to support error recovery. In real-robot experiments, its success rate was 77%, more than double the 37.1% of the strongest plan-then-act baseline.
arXiv cs.RO ·