NOITOM ROBOTICS NR Ego-Vision Data Acquisition Device
The NOITOM ROBOTICS NR Ego-Vision Data Acquisition Device is a multimodal perception and computing platform designed for embodied AI, robot learning, and first-person data collection. It captures human–object–environment interactions and combines visual, inertial, tracking, and other sensor data into synchronized datasets for AI training and robotic interaction.
The device combines dual fisheye RGB cameras with a 64 mm baseline, 50 MP sensors, 3K effective imaging resolution, HDR, and a 200° field of view. The cameras are synchronized with other acquisition sensors for consistent multimodal data capture.
Motion and Spatial Perception
Four monochrome cameras provide head movement tracking, depth recognition, and hand motion capture, while the system also supports real-time VSLAM pose output. This makes it suitable for collecting human motion and interaction data for robotics research.
Synchronized Multimodal Data
The platform supports hardware-triggered multimodal synchronization and ≤1 ms multi-device synchronization through its 2.4G private protocol. This enables multiple sensing devices to work together when building high-quality datasets for embodied AI and robot learning.
Edge Computing and Flexible Connectivity
Powered by the Qualcomm XR2 Gen 1 platform with 8 GB LPDDR RAM and 256 GB UFS storage, the device supports Wi-Fi 6/6E, Bluetooth 5.2, removable storage, and USB-C connectivity. A power bank can extend operation to up to 8 hours for longer data collection sessions.
