prophysics AG is your partner in the DACH region for professional motion capture systems in humanoid robotics. We offer full-body kinematics and precise finger tracking, from teleoperation to AI training.<\/p>\n
Humanoid robots require high-quality kinematic data. This data covers the entire body, including the hands, and must be synchronised in real time. It forms the basis for teleoperation, imitation learning and whole-body control.<\/p>\n
The robot\u2019s AI learns from examples of human movement. Precise whole-body and finger data form the basis for the \u2018learning from demonstration\u2019 approach and generalisable foundation models.<\/p>\n<\/div>\n
With a latency of ~20 ms, a person wearing a MoCap suit controls a robot directly. Seamless integration is achieved via UDP, DDS and the SDK in ROS\/ROS 2. Each session simultaneously generates valuable training data.<\/p>\n<\/div>\n
Whole-Body Motion Retargeting transfers human movement patterns directly to the joint and motion models of humanoid robots \u2013 from low-level control right up to the high-level AI stack.<\/p>\n<\/div>\n<\/div>\n
Manipulation and grasping tasks are the biggest challenge in robotics. EMF-based finger tracking with 25 degrees of freedom (DOF) can achieve what camera systems cannot when occlusion occurs.<\/p>\n<\/div>\n
High-quality, scalable training data is available for learning from demonstration and imitation learning, as well as for the validation of AI-based motion and decision-making models.<\/p>\n<\/div>\n
Natural and precise interaction data are essential for research into human-robot interaction and embodied AI. The whole body and hands are captured together in a synchronised session.<\/p>\n<\/div>\n<\/div>\n