Research map / Vision-language and multimodal / Robot action and navigation
TCoT: Trajectory Chain-of-Thoughts for Robotic Manipulation with Failure Recovery in Vision-Language-Action Model
Xiang Li, Ya-Li Li, Yuan Wang, Huaqiang Wang, Shengjin Wang
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 |
|---|---|---|---|
| add global trajectory planning (ablation) | OpenVLA · LIBERO average | success rate: 0.765 -> 0.786 | Improved |
| add local trajectory planning (ablation) | OpenVLA · LIBERO average | success rate: 0.765 -> 0.773 | Improved |
| add global and local planning plus failure detection and recovery (GLSR) (ablation) | OpenVLA · LIBERO average | success rate: 0.786 -> 0.829 | Improved |
| replace OpenVLA with TCoT on long-horizon suite | OpenVLA · LIBERO-Long | success rate: 0.537 -> 0.668 | Improved |
| train TCoT multi-task instead of single-task | TCoT · LIBERO average | success rate: 0.829 -> 0.833 | Improved |
| train OpenVLA multi-task instead of single-task | OpenVLA · LIBERO average | success rate: 0.765 -> 0.710 | Hurt |
| replace OpenVLA with TCoT on real-world tasks | OpenVLA · AIRBOT single-task average | success rate: 0.39 -> 0.67 | Improved |
| train TCoT multi-task instead of single-task | TCoT · AIRBOT real-world | success rate: 0.67 -> 0.75 | Improved |
| train OpenVLA multi-task instead of single-task | OpenVLA · AIRBOT real-world | success rate: 0.39 -> 0.43 | Improved |
| add TCoT on OpenVLA-OFT continuous-action backbone | OpenVLA-OFT · LIBERO-Long | success rate: 91.1 -> 93.5 | Improved |
| add TCoT on OpenVLA-OFT continuous-action backbone | OpenVLA-OFT · LIBERO average | success rate: 95.4 -> 96.2 | Improved |
| add TCoT on OpenVLA-OFT | OpenVLA-OFT · LIBERO-Goal | success rate: 95.5 -> 95.0 | No clear effect |
Later papers that compared with this method
1 later accepted paper reports TCoT as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Cut to the Chase: Training-free Multimodal Summarization via Chain-of-Events (CVPR 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CoT-VLA — CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models (CVPR 2025) · better on 3 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- AHA: A Vision-Language-Model for Detecting and Reasoning Over Failures in Robotic Manipulation (ICLR 2025 · similarity 0.91)
- RoboProcessBench: Benchmarking Process-Aware Understanding in Visual-Language Robotic Manipulation (NeurIPS 2026 · similarity 0.89)
- CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models (CVPR 2025 · similarity 0.89)
- Failing Forward: Adaptive Failure-Informed Learning for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.89)
- SAFE: Multitask Failure Detection for Vision-Language-Action Models (NeurIPS 2025 · similarity 0.88)
- RePO-VLA: Recovery-Driven Policy Optimization for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.88)
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.
