Teaching Robots Theory of Mind to Boost Collaboration
This innovative framework, named HUMAC, was presented at the IEEE International Conference on Robotics and Automation (ICRA 2025), held in Atlanta, Georgia, from May 19-23, 2025. The system allows a single human operator to guide multiple robots through brief, targeted interventions, embedding human-like teamwork strategies into their algorithms.
"Humans start to develop the skill of Theory of Mind around age four," said Boyuan Chen, the Dickinson Family Assistant Professor at Duke University. "It allows us to interpret and predict others' intentions, allowing collaboration to emerge. This is an essential capability that our current robots are missing to allow them to work as a team with other robots and humans. We designed HUMAC to help robots learn from how humans think and coordinate in an efficient way."
Traditional methods for teaching robotic collaboration, such as reinforcement learning and imitation learning, require either extensive trial-and-error or costly expert demonstrations, making them impractical for many real-world applications. In contrast, HUMAC uses a more streamlined approach, enabling a single human to take control of individual robots at critical moments, much like a coach directing a team on the field.
The research team tested this system in a dynamic hide-and-seek scenario, where seeker robots aimed to catch faster hider robots within a confined, obstacle-filled arena. With just 40 minutes of human-guided training, the robots achieved a success rate of 84% in simulations and 80% in physical tests-dramatically higher than the 36% success rate of non-cooperative seekers.
"We observed robots starting to behave like genuine teammates," said Zhengran Ji, the lead student author and graduate student in Chen's lab. "They predicted each other's movements and coordinated naturally, without explicit commands."
Looking ahead, the researchers plan to expand HUMAC to larger, more complex robot teams, potentially revolutionizing applications such as disaster response and coordinated drone swarms.
"AI is not just a tool for humans, it's a teammate," Chen said. "Just as humans evolved to collaborate, AI will become more adaptive to work alongside each other and with us. HUMAC is a step toward that future."
Research Report:Enabling Multi-Robot Collaboration from Single-Human Guidance
Related Links
General Robotics Lab
All about the robots on Earth and beyond!
