Researchers Enable Humanoid Robots to Distinguish Self from Others Using Proprioceptive-Visual Correspondence

A new research paper demonstrates that humanoid robots can learn to distinguish self from others through proprioceptive-visual correspondence, without identity labels or kinematic models. Once established, this distinction bootstraps a predictive self-model that maps joint configurations to 3D body occupancy. In multi-agent scenarios involving humans or morphologically identical robots, the system reliably identifies itself, learns a 3D self-model, and supports downstream tasks including target reaching, collision-aware motion planning, and human-to-robot motion retargeting.