Moxian Technology has launched a full series of data-collection gloves aimed at the embodied-intelligence sector, according to Chinese outlet D1EV. The six gloves were unveiled during IROS 2026, the IEEE/RSJ International Conference on Intelligent Robots and Systems, and are designed to capture tactile, bending and motion information from the hand for continuous data collection in real operating scenarios.
High-density sensing and precise tracking
The gloves offer high-density tactile perception across the full palm, with 360 to 860 tactile points per glove and a spatial resolution of 2mm. A local rigid-support composite structure suppresses stretching, wrinkling and finger-bending interference, keeping point repeatability and consistency within 10% of full scale (FS).
Each glove is paired with a high-precision electromagnetic positioning sensor that enables pose reconstruction across 25 degrees of freedom for the whole hand, and six degrees of freedom tracking for each fingertip, with positioning accuracy of less than 0.2mm. The system supports synchronized tactile and pose sampling at 120Hz, with a wired latency of 8ms and a wireless latency of 30ms.
Applications in embodied intelligence
The product line is primarily used for data collection, operation learning, dexterous-hand teleoperation and multimodal dataset construction in embodied intelligence, as well as human-machine interaction and scientific research. Moxian said it offers product selection, sample testing and customized development services for different applications. The company demonstrated the gloves at booth 329 during the exhibition, showing real-time collection and conversion of multimodal information.
According to D1EV, the release responds to growing demand for high-quality training data in the embodied-intelligence field, where robots increasingly need rich tactile and motion data to learn manipulation skills. The gloves are positioned as a tool for developers and researchers building multimodal datasets for robot learning.
