This repository is a curated, community-driven list of academic papers, code, and resources focused on robot manipulation. It serves researchers, engineers, and practitioners in robotics seeking to stay abreast of the latest advancements in areas like grasping, motion planning, and vision-language models for robotic control. The primary benefit is a comprehensive, organized overview of a rapidly evolving field.

How It Works

The repository categorizes papers by sub-fields within robot manipulation, such as Grasping (Rectangle-based, 6-DoF, Language-Driven), Representation Learning (Visual, Multimodal), and specific techniques like Diffusion Policy Learning and Vision-Language Action Models. Each entry typically includes the paper title, venue, date, and a link to its code or project page, facilitating quick access to relevant research.

Quick Start & Requirements

This repository is a curated list and does not require installation or execution. It serves as a reference guide.

Highlighted Details

• Extensive coverage of recent (2023-2025) research papers across numerous robot manipulation sub-domains.
• Categorization includes emerging areas like Diffusion Policy Learning, Vision-Language Action Models, and 3D Representation for Manipulation.
• Many entries link directly to GitHub repositories for associated code, enabling practical exploration.
• Includes lists of relevant benchmarks and datasets for evaluating manipulation systems.

Maintenance & Community

The repository is actively maintained and welcomes community contributions via pull requests or issues.

Licensing & Compatibility

The repository itself is licensed under an unspecified license, but it curates links to external research papers and code, each with its own licensing.

Limitations & Caveats

As a curated list, the depth of information for each paper is limited to what is typically found in a paper's abstract and associated repository. The "latest" papers are often pre-prints or recently accepted publications, meaning code may be experimental or not yet widely validated.