【具身智能/自主导航】相关开源项目代码、论文收集

1. 代码集合

• Awesome-LLM-Robotics: https://github.com/GT-RIPL/Awesome-LLM-Robotics
• Everything-LLMs-And-Robotics: https://github.com/jrin771/Everything-LLMs-And-Robotics
• Awesome LLM-Powered Agent: https://github.com/hyp1231/awesome-llm-powered-agent
• awesome-embodied-vision: https://github.com/ChanganVR/awesome-embodied-vision

2. 论文集合

• CoRL | OpenReview: https://openreview.net/group?id=robot-learning.org/CoRL
• Embodied AI Workshop (CVPR): https://embodied-ai.org/
• Springer: Autonomous Robots: https://link.springer.com/journal/10514/collections
• Habitat: A Platform for Embodied AI Research: https://aihabitat.org/

3. 自主导航强相关项目

• LM-Nav: Robotic Navigation with Large Pre-Trained Models: https://sites.google.com/view/lmnav/home
• Visual Language Maps for Robot Navigation: https://vlmaps.github.io/
• ConceptFusion: Open-set Multimodal 3D Mapping: https://concept-fusion.github.io/
• ESC: Exploration with Soft Commonsense Constraints for Zero-shot Object Navigation: https://sites.google.com/ucsc.edu/escnav/home
• ViNG: Learning Open-World Navigation with Visual Goals: https://sites.google.com/view/ving-robot/
• RECON: Learning to Explore the Real World with a Ground Robot：https://sites.google.com/view/recon-robot
• ViKiNG: Vision-Based Kilometer-Scale Navigation with Geographic Hints: https://sites.google.com/view/viking-release
• General Navigation Models: https://general-navigation-models.github.io/

4. 其他相关项目

• CLIP-Fields: Weakly Supervised Semantic Fields for Robotic Memory: https://mahis.life/clip-fields/
• CoWs on Pasture: Baselines and Benchmarks for Language-Driven Zero-Shot Object Navigation: https://cow.cs.columbia.edu/
• LERF: Language Embedded Radiance Fields: https://www.lerf.io/- Audio Visual Language Maps for Robot Navigation": https://avlmaps.github.io/
• Simple but Effective: CLIP Embeddings for Embodied AI: https://github.com/allenai/embodied-clip
• Semantic Abstraction: Open-World 3D Scene Understanding from 2D Vision-Language Models: https://semantic-abstraction.cs.columbia.edu/

更多

• 欢迎留言补充
• 个人认为具备开源代码、论文、网页的项目更容易收获影响力，特别网页是一个非常好的展示形式。