Human Centred Robot Learning

Project overview

As robots move from factories into homes, hospitals, and public spaces, their ability to adapt to human needs and environments becomes crucial. This project explores how robots can learn context-specific tasks directly from human trainers—including people with no programming background—while keeping the teaching process intuitive, engaging, and effective.

Grounded in educational theory and human–computer interaction, our work focuses on making robot learning a genuinely human-centred process. We investigate:

• Training humans to train robots—how to help novice users give effective demonstrations.
• Reducing disengagement—minimising fatigue and loss of interest during repetitive teaching tasks.
• Designing engaging interactions—using gamification and instructional design to sustain motivation and improve learning outcomes.

The project combines Learning from Demonstration (LfD) methods with user interface innovation, mixed reality tools, and novel interaction strategies to make robot training accessible, efficient, and even enjoyable.
This three-year initiative, funded by the Australian Research Council (2021-2024) DE210100858.

Current directions

Our research explores both the robot’s perspective—improving how machines interpret and learn from human input—and the human’s perspective—making the teaching process intuitive, rewarding, and less demanding. Recent studies have:

• Developed metrics for quantifying demonstration quality beyond just task success.
• Introduced mixed reality interfaces that help users visualise and refine their robot training.
• Applied competitive and collaborative interaction designs to influence both robot performance and human perceptions.
• Designed user interface interventions that improve the efficiency of LfD sessions.

Publications

• Beyond Success: Quantifying Demonstration Quality in Learning from Demonstration – Bilal, M., Lipovetzky, N., Oetomo, D., & Johal, W. – IROS 2024.
• Mr.LfD: A Mixed Reality Interface for Robot Learning from Demonstration – Chen, J., Chacon, A., Bilal, M., Zhou, Q., & Johal, W. – OzCHI 2024.
• Improving Robot Learning from Demonstration via Competitive Interactions and User Interface Interventions – Phaijit, O., Sammut, C., & Johal, W. – ACRA 2024.
• User Interface Interventions for Improving Robot Learning from Demonstration – Phaijit, O., Sammut, C., & Johal, W. – HAI 2023.
• Let’s Compete! The Influence of Human-Agent Competition and Collaboration on Agent Learning and Human Perception – Phaijit, O., Sammut, C., & Johal, W. – HAI 2022.
• A Taxonomy of Functional Augmented Reality for Human-Robot Interaction – Phaijit, O., Obaid, M., Sammut, C., & Johal, W. – HRI 2022.