Data as an instrument for more efficient collaboration with robots
summary
Lab builds and tests stiffness-controllable actuators that let cobots modulate joint compliance, combines vision, IMUs and EMG to predict human motion and emotion, and uses inverse optimal control to infer humans’ task strategies. Mechanical impedance is estimated with model-based observers rather than direct sensing, and a modified dynamic time warping collision detector separates intentional contact from harm. A factory pilot guides operators through 20 similar assemblies, detects loss of focus, and reduces human error while logging rehabilitation- and production-grade datasets.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:18Robotics history
- 3:50Industry 4.0 applications
- 6:41Industry 5.0 transition
- 7:43Human-robot safety
- 9:09ETF Robotics Lab
- 10:13Cobot research equipment
- 11:17Cobot actuators & motion learning
- 13:45Human motion & emotion prediction
- 15:31Robot impedance estimation
- 16:13Collision detection algorithms
- 17:19Human-centered assembly
- 18:45Worker mental workload
- 19:53Assembly innovation demo
- 23:42Research publications
- 25:29Industry & healthcare partners
- 26:30Q&A
related talks
18mRobotics: why now?
09mTask-Consistent Signaling Motions for Improved Understanding in Human-Robot Interaction and ...
22mSeeing, Sensing and Recognizing Laban Movement Qualities
10mEnabling Robot-assisted Motion Capture with Human Scale Tracking Optimization
17mOpen Source Robotics: Advances in Teleoperation for Dexterous Manipulation
22mUsing data to improve physical therapy for people with neurological diseases
08mHuman–AI Collab for Reliable Automotive Control Systems
47mDr. Simon Müller | AI Track | Compass AI & Data Summit 2025
