Human-Inspired Principles for Robotic Dexterous Manipulation: What are we Missing?

Date: June 5th, 2026
Time: 09:00 AM
Location: TBD

Abstract: Robotic manipulation is rapidly advancing, yet robots remain far from matching humans in contact-rich, highly dexterous physical interaction. Humans excel at such tasks despite slow actuators, noisy sensing, and neural delays—raising the question: what are we missing in robots? Recent progress in robot learning (reinforcement learning, diffusion policies, foundational models) has improved generalization in motion planning and control, while model-based approaches (MPC, optimal and whole-body control) remain powerful in well-structured domains. Still, both paradigms struggle to scale to complex contact dynamics. At the same time, robotics hardware—actuation and sensing—lags behind human capability. Muscles can flexibly shift from highly compliant to stiff behaviors, while touch and vision jointly guide human interaction with the physical world. Can biological principles inspire new designs and control strategies?

Mission Statement: This workshop brings together experts in robotics, biomechanics, and motor neuroscience to critically assess current progress and identify key barriers. Through invited talks, contributed presentations, and interactive discussions, we will explore open questions at the intersection of hardware, control, and learning. The goal is to chart a roadmap for advancing robot physical interaction, foster cross-disciplinary collaboration, and engage a diverse community of researchers at all career stages.