Overview

ROS4HRI (Robot Operating System for Human-Robot Interaction) is the standard ROS API for representing and managing information about humans in a robot’s environment. Officially accepted as REP-155, it provides a common language for social robots to perceive, interpret, and respond to human signals, independent of the underlying sensors or algorithms.

Objectives

• Standardization: Establish a unified ROS API (REP-155) for human representation (faces, bodies, voices, and persistent person entities).
• Interoperability: Enable different HRI components (person trackers, face recognizers, gesture detectors) to work together seamlessly.
• Reusability: Promote the development of sensor-agnostic and algorithm-agnostic HRI software tools.
• Reproducibility: Facilitate benchmarking and comparative evaluation of HRI algorithms through standardized interfaces.

Technology

• REP-155 Specification: The core standard defining naming conventions and message structures for social robotics.
• hri_msgs: Standardized ROS messages for human-related data such as skeletal joints, 2D/3D regions of interest, and engagement levels.
• libhri: High-level C++ and Python libraries designed to simplify the aggregation and access of human-related signals for developers.
• Multimodal Integration: Support for fusing data from vision, audio, and depth sensors into persistent, unique person IDs.
• Visualization Tools: Specialized RViz plugins for real-time visualization of human skeletons, gaze, and social space.

Impact

ROS4HRI is transforming social robotics development by reducing redundant custom implementations and enabling modular HRI software stacks. It is a foundational component in numerous international research projects ranging from autism therapy (https://socialminds.iiia.csic.es/projects/emorobcare/) to human-centered industrial collaboration. By making HRI software shareable across different robot platforms, it significantly accelerates the transition from university research to real-world social impact.

Get Involved

• GitHub: Contribute to the core libraries and perception modules on our GitHub organization.
• Documentation: Explore the complete specification, tutorials, and getting started guides at ros4hri.github.io.
• Community: Join the discussion on the ROS Discourse HRI category to help shape the future of the standard.

Publications highlights

ROS for Human-Robot Interaction

Lemaignan, S. , Mohamed, Y. , Séverin Lemaignan

2021 - Proceedings of the 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems DOI: 10.1109/IROS51168.2021.9636816 Social Robotics

Probabilistic Fusion of Persons' Body Features: The Mr. Potato Algorithm

Lorenzo Ferrini , Séverin Lemaignan

2024 - DOI: 10.1145/3610977.3637479 Social Robotics

Hands-on: From Zero to an Interactive Social Robot Using ROS4HRI and LLMs

Ferran Gebelli , Lorenzo Ferrini , Luka Juričić , Raquel Ros , Sara Cooper , Séverin Lemaignan

2025 - DOI: 10.1109/HRI61500.2025.10974214 Social Robotics

Socially Pertinent Robots in Gerontological Healthcare

Alessandro Conti , Alex Auternaud , Angus Addlesee , Anne-Sophie Rigaud , Ariel Rado , Bin Ren , Chris Reinke , Christian Dondrup , Cigdem Beyan , Cyril Liotard , Daniel Hernández García , Elisa Ricci , Federica Arrigoni , Florian Gras , Francesco Ferro , Francesco Tonini , Imad Kimouche , Lauriane Blavette , Luca Marchionni , Luis Gomez Camara , Maribel Pino , Marta Romeo , Matthieu Py , Michal Polic , Mordehay Moradi , Moreno D’Incà , Nancie Gunson , Nicolas Turro , Nicu Sebe , Ohad Cohen , Oliver Lemon , Paolo Rota , Pinchas Tandeitnik , Radu Horaud , Sharon Gannot , Soraya Arias , Sébastien Dacunha , Séverin Lemaignan , Timothée Wintz , Tomas Pajdla , Weronika Sieińska , Xavier Alameda-Pineda , Yanchao Yu , Yoav Ellinson

2025 - International Journal of Social Robotics DOI: 10.1007/s12369-025-01330-6 Social Robotics