Research map / Vision-language and multimodal / Robot action and navigation
ODYSSEY: Open-World Quadrupeds Exploration and Manipulation for Long-Horizon Tasks
Kaijun Wang, Liqin Lu, Mingyu Liu, Jianuo Jiang, Zeju Li, Bolin Zhang, Wancai Zheng, Xinyi Yu, Hao Chen, Chunhua Shen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace multi-camera end-to-end PerAct with single-egocentric-camera VLM planner | ODYSSEY · ARNOLD novel splits | success rate (%): 25.70 -> 45.24 (PickUpObject), 41.32 -> 79.50 (CloseCabinet) | Improved |
| replace PerAct with ODYSSEY on seen PickUpObject | ODYSSEY · ARNOLD PickUpObject (seen) | success rate (%): 94.03 -> 60.45 | Hurt |
| replace PerAct with ODYSSEY on seen ReorientObject, OpenCabinet, CloseCabinet | ODYSSEY · ARNOLD seen splits | success rate (%): 19.48 -> 51.32 (ReorientObject) | Improved |
| use single unified policy with terrain-invariant end-effector sampling instead of RoboDuet dual-policy | ODYSSEY policy · Moving base tracking | eω error: 60.59 -> 0.79 | Improved |
| same swap | ODYSSEY policy · Moving ex error | ex: 9.70 -> 0.36 | Improved |
| same swap | ODYSSEY policy · Standing ey error | ey: 0.34 -> 2.69 | Hurt |
| same swap | ODYSSEY policy · End-effector pose tracking | Dpos/Dori: 11.08 -> 11.48 (Dpos, standing) | No clear effect |
Closest papers by meaning
- Instruction-Augmented Long-Horizon Planning: Embedding Grounding Mechanisms in Embodied Mobile Manipulation (AAAI 2025 · similarity 0.90)
- SPIN: Simultaneous Perception Interaction and Navigation (CVPR 2024 · similarity 0.89)
- MoManipVLA: Transferring Vision-language-action Models for General Mobile Manipulation (CVPR 2025 · similarity 0.88)
- Humanoid Horizon: Extending Task Horizon in Whole-Body Loco-Manipulation via Parallel Training, Dynamic Starting, and Reward Gating (NeurIPS 2026 · similarity 0.87)
- OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints (CVPR 2025 · similarity 0.87)
- RoboCerebra: A Large-scale Benchmark for Long-horizon Robotic Manipulation Evaluation (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
