Skip to content

Data Collection Overview ​

Data collection is where the NexCore workflow begins. The platform supports two collection methods, plus a third data source that requires no dedicated collection sessions:

MethodDevicesWhat is capturedStorage formatLearn more
Teleoperation data collectionStation Lite (dual-arm workstation), MOS (wheeled dual-arm robot)An operator teleoperates the robot, recording robot state, actions, and multiple camera streamsHDF5Teleoperation Data Collection
Egocentric data collectionRobot-free Ego, Robot-free UMIA person performs the demonstration directly, recording first-person stereo video, IMU, and other raw sensor dataRAWEgocentric Data Collection
Continuous operation dataRobots running in productionThe full working day (teleoperation, autonomous, takeover, idle, and failure) flows back to the platform5-minute archives—

Choosing a Method ​

ConsiderationTeleoperationEgocentric
Consistency with the target robotHigh: actions are produced directly on the target embodimentRequires pipeline processing and retargeting
Collection throughputLimited by operator skill and the number of robotsHigher: people demonstrate directly
Scene coverageThe robot must be brought to the sceneLightweight devices reach more real-world environments
Downstream processingCan be curated directlyMust go through a data pipeline first

Both kinds of data can be combined in one project, for example egocentric data to broaden scene coverage and teleoperation data to keep the embodiment consistent.

Collection Workflow ​

text
Register device → Create collection task → Collector claims task and records → Upload → Integrity check → Data assets
StepDescriptionReference
Register deviceAdd the device model and SN to the device inventoryDevice Management
Create collection taskSelect model, select SN, select upload mode, enter task detailsCollection Tasks
Collect and uploadThe collector runs the task; data uploads according to the upload modeData Upload
Validate and ingestData is assigned a data ID once channels, timestamps, and package counts are consistentData Assets

Collection Best Practices ​

  • Start small: Record 10 to 20 episodes, curate them, and run a short training job to validate the pipeline before collecting at scale.
  • Keep demonstrations consistent: Within a task, keep the starting pose, object placement range, and pace as consistent as possible.
  • Cover variation deliberately: Within that consistency, intentionally vary object positions, lighting, and backgrounds.
  • Keep failures: Failed demonstrations don't need to be deleted. Flag them during curation, because failure data is valuable too.
  • Document the setup: Record the scene layout, object specifications, and similar details in the task notes for later traceability.

NexCore is evolving rapidly; the latest version prevails.NexCore Docs · Embodied AI Skill Learning Platform by Lumos Robotics