Human co-manipulation via motion and haptic information for physical human-robotic collaboration

• Understanding human co-manipulation via motion and haptic information can enable future physical human-robotic collaborations.¹
• Motion and haptics inform shared-control and assistive interfaces.¹
• No neural recording or decoding; implementation path is human factors and robotics; tangential to BCI.^{1 1}

Weekly enrichment (2026-07-20)

• The underlying study, from Brigham Young University’s Robotics and Dynamics Laboratory (Shaw, Salmon, and Killpack), appeared in Frontiers in Neurorobotics in 2025 and dissects four sub-components of human co-manipulation to inform future physical human-robot collaboration.^{2 3}
• Data came from 16 sessions with 16 different three-person teams who moved a large instrumented object using haptic communication only, with visual and auditory channels deliberately removed via a virtual reality setup.^{2 3}
• The co-manipulated object was a 1.3 m by 0.5 m table weighing 25.3 kg, fitted with four ATI Mini45 force/torque sensors and six HTC Vive trackers; force, torque, position, and velocity were recorded at 200 Hz.^{2 3}
• Teams were tested under three leadership configurations — leader-leader (LL, both informed), leader-follower (LF), and leader with two followers (LFF) — to probe how shared goal knowledge and added followers affect haptic-only coordination.^{2 3}
• The analysis defined a method for detecting the transition from a static/rest state to an active state, so a robot teammate could learn when to wait versus when to move.^{2 3}
• Across the six rigid-body degrees of freedom, the authors compared candidate signals (force, acceleration, velocity, etc.) to identify which best predict the team’s intended motion primitive, and produced a ranked list of which DOFs are easy versus hard to convey through haptics alone.^{2 3}
• The work is a human-factors and robotics study with no neural recording or decoding, so its relevance to BCI is indirect — it informs shared-control and assistive interface design rather than direct brain-based control.^{2 3}

Footnotes

1. https://news.google.com/rss/articles/CBMilgFBVV95cUxPV2NJSWtIcUJmaDVqRVhFRy1rb3ZKcnI1c0FKZDkxcVNhcXdlS1cwQWRScDRUTEFjLWpqc19ybTNiSkEtN1pzWlhJUEZSWTktdS1GOU4tcG4wTkJvMEtzeU5GVHliNVlRMVQ4WDVPQVgxNmc5WDFNYWFBeFF3NjNxdExaNENJMEljajVZVlVKVERuR1B3Z1E?oc=5 ↩ ↩² ↩³ ↩⁴

2. https://www.frontiersin.org/journals/neurorobotics/articles/10.3389/fnbot.2025.1480399/full ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

3. https://pmc.ncbi.nlm.nih.gov/articles/PMC12222233/ ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷