Aligned multimodal records
Capture vision, tactile/contact data, IMU, 6-DoF pose and gripper state around one task timeline.
Multimodal Data Collection for Embodied AI
Collect real-world robot learning data with vision, tactile signals, IMU, pose and gripper state.
T-Grip Suite and OmniView Suite help robotics teams build synchronized manipulation datasets for VLA, imitation learning, diffusion policy, teleoperation and in-the-wild robot teaching.
4+Sensor modalities
SDKSecondary development
$19.1kCombined package
RGB-D
IMU
Pose
Tactile
Gripper
Timestamped streams
Robot-ready actions
Model-training datasets
The Bottleneck
Robot models need physical interaction data, not only videos.
For contact-rich tasks such as insertion, alignment, cable handling and delicate grasping, visual data alone often misses force, touch and micro-motion cues. Our system is built to capture those signals in a developer-friendly workflow.
Lower-cost real-world teaching
Use handheld and wearable workflows to collect data before building expensive full robot workcells.
Open development direction
Package delivery can include drivers, SDK, calibration tools, export utilities and retargeting interfaces.
Products
Two suites, one complete embodied-data workflow.
T-Grip focuses on manipulation and contact data. OmniView focuses on first-person wearable context. Together they cover both task execution and operator environment.
Manipulation Data
T-Grip Suite
A dexterous data collection gripper for handheld capture, teleoperation and robot-teaching workflows.
- Joint motor, handle, IMU, RGB-D camera, tactile sensor and finger sleeve architecture.
- Designed to record TCP pose, gripper state and synchronized manipulation signals.
- Suitable for insertion, alignment, grasping, assembly and tactile-vision learning.
- Includes secondary-development software package development.
Wearable Perception
OmniView Suite
A wearable data collection system for first-person, mobile, bimanual and in-the-wild operation data.
- Compute board, battery, contact pad, stretch strap, camera and IMU architecture.
- Designed for scene capture, wearable sensing and operator-context data.
- Suitable for agriculture, mobile manipulation, bimanual teaching and field collection.
- Includes secondary-development software package development.
Workflow
From synchronized acquisition to robot-learning datasets.
The pipeline is designed around timestamped sensor streams, software alignment, visual stitching, SLAM-style scene understanding and final data output for downstream model training.
System Direction
Built as data infrastructure for robotics teams.
Exact camera, IMU, tactile and synchronization specifications can be confirmed by customer configuration and R&D validation.
Data modalities
RGB-D / camera data, IMU, tactile or contact-related signals, gripper state and 6-DoF pose-oriented data.
Dataset structure
Timestamped streams, TCP pose, Delta Pose conversion, image/video records and robot-learning dataset export.
Developer support
Drivers, calibration tools, data export utilities, SDK and retargeting interfaces can be provided by project scope.
Target workflows
VLA, imitation learning, diffusion policy, teleoperation, manipulation data collection and robot policy training.
Pricing
A cost-effective entry point for multimodal embodied-data collection.
T-Grip Suite
¥59,800 ≈ $8,800 USD
Dexterous data collection gripper suite with secondary-development software package development.
OmniView Suite
¥75,800 ≈ $11,150 USD
Wearable perception and data collection suite with secondary-development software package development.
Combined Package
¥129,800 ≈ $19,100 USD
T-Grip Suite + OmniView Suite for a complete handheld, wearable and multimodal collection workflow.
Best-fit Customers
For teams turning physical interaction into model-ready data.
Embodied AI startups
Robot learning labs
Humanoid robot teams
Agricultural robotics
Precision assembly automation
VLA and imitation-learning teams
Contact
Tell us your target task, robot platform and data format.
We will suggest a suitable T-Grip / OmniView configuration, demo plan and development support scope.
WeChat: Ustartechco