reMap is an open-source framework designed to enable robots to represent and interact with spatially-grounded semantics. It integrates both spatial and symbolic data to facilitate human-robot interaction (HRI) in dynamic environments. The system leverages a VDB grid to efficiently store and manage semantic entities and their spatial relationships.

This project is implemented as a ROS 2-based package, allowing seamless integration into modern robotic applications. It supports querying through SPARQL-based semantic web APIs, making it highly adaptable to various knowledge representation frameworks.

• Voxel-based spatial representation: Efficient handling of sparse and dynamic 3D information.
• Semantic integration: Links spatial data to symbolic concepts in an RDF-based knowledge base.
• Modular ROS 2 architecture: Includes plugins to process different types of semantic data.
• Real-time querying: Enables retrieval of spatially-grounded information via SPARQL.
• Extensible framework: Designed for easy adaptation and expansion for diverse robotic applications.

Anyone is welcome to contribute to this project. Feel free to open an issue or submit a pull request to the specific package you want to improve.

This work is part of the H2020 PERSEO project (grant no. 955778). Special thanks to PAL Robotics for supporting development and testing. The authors of the organization profile pictures are Vossman and M. W. Toews, and is available under CC BY-SA 2.5 license.