[Free Seminar] Humanoid Robot Learning Data Seminar

For a humanoid robot to walk, grasp objects, and perform tasks like a human, it first needs data that teaches it how humans move. As research in Physical AI and humanoid robotics accelerates, many labs and companies are facing the same questions:

• How do you capture motion data precise enough to use directly for robot learning?

• How do you integrate motion capture data with data from other sensors?

• How do you connect that data to real robot learning, such as imitation learning and reinforcement learning?

This seminar answers these questions with live demonstrations on real equipment.

Program

Part 1 | Capturing Motion Data for Robot Learning

Setting up an optical motion capture environment and collecting high-quality motion data

The quality of robot learning data is decided at the capture stage. Part 1 introduces how to collect robot learning data using optical motion capture. See everything on site, from setting up the capture environment and tuning for higher accuracy to the actual capture process.

Part 2 | From Captured Data to Robot Learning

Sensor fusion data integration and the humanoid training process

Raw captured data can't be fed straight into robot learning. Part 2 covers the sensor fusion process of synchronizing and integrating motion capture data with data from other sensors, and walks through the full pipeline that connects it to humanoid robot learning. The session includes a live demo of a humanoid learning data integration platform.

Event Details

Date	Friday, October 16, 2026, 16:00–18:00 (KST)
Venue	Motion Capture Studio, Room 507, Building 2, Seongsu SK V1 Center, 31 Gwangnaru-ro 8-gil, Seongdong-gu, Seoul, Korea
For	Researchers and engineers working on humanoid and robot learning data
Admission	Free
Register	https://forms.gle/ZKM3WJVPU4RYciLc7

Register Now