iCub-HRI: A Software Framework for Complex Human-Robot Interaction Scenarios on the iCub Humanoid Robot

Area of Science:

• Robotics
• Artificial Intelligence
• Human-Robot Interaction

Background:

• Developing human-like behavior in humanoid robots necessitates integrating diverse open-source components and scalable cognitive architectures.
• The iCub platform requires robust software frameworks for advanced interaction capabilities.

Purpose of the Study:

• To introduce the iCub-HRI library, a C++ framework with Java, Python, and Matlab bindings.
• To provide convenience wrappers for perception, manipulation, and social interaction components.
• To enable extensions for new components and rapid prototyping in human-robot interaction.

Main Methods:

• Developed a C++ library (iCub-HRI) with bindings for Java, Python, and Matlab.
• Integrated components for object recognition, agent tracking, speech recognition, touch detection, motor actions, speech synthesis, and joint attention.
• Created modules for knowledge acquisition and action recognition.

Main Results:

• The iCub-HRI library successfully supports complex human-robot interaction scenarios.
• Demonstrated language acquisition, goal-oriented behavior, and narrative expression in a humanoid robot.
• A tutorial is provided for reproducing a subset of the interaction.

Conclusions:

• The iCub-HRI library offers a flexible and extensible platform for humanoid robot research.
• It is suitable for both novice and expert users within the iCub ecosystem.
• The library is expected to foster widespread adoption and advance human-robot interaction research.