Exo-Data Standard

To address these challenges, we propose the Exo-Data Standard, a data service framework designed for multi-center research environments. Beyond simple data collection, this framework standardizes and integrates heterogeneous multimodal datasets—including human kinematics, kinetics, and bio-signals, as well as robot state information and assistive torque data—under a unified protocol. By ensuring interoperability across diverse hardware platforms and experimental environments, the framework establishes a scalable data foundation applicable to real-world clinical and daily-life scenarios. The high-quality datasets accumulated through this platform will serve as a foundation for training next-generation AI models capable of motion intention recognition, gait pathology analysis, and personalized control optimization, thereby accelerating the development of Physical AI systems that enable seamless human–robot interaction.