ZIXITECHROBOTICS · FLEXIBLE SENSING
English中文
Technical articles2 min read

From hand pose to useful robot data

Choose the signals first. Then follow one task from hand motion to a recording you can review.

A hand model for motion mapping

The essentials

  • Motion tracking, pressure measurement and haptic feedback serve different roles.
  • Validate the target-hand mapping before collecting a dataset.
  • Record the signals, task events and software configuration together.
On this page

Choose what the operator needs to control

A finger-level task needs more than a moving hand model. Identify which fingers matter, where contact should be measured and whether the wearer needs a vibration or resistance cue.

MOTION

Track the movement

Capture hand pose and turn it into the target hand’s commands.

CONTACT & FEEDBACK

Understand the interaction

Measure contact signals, or send cues back to the wearer. A pressure sensor measures; an actuator delivers feedback.

Fit the glove and check the mapping

Record a comfortable open hand, a relaxed bend and the movements used in the task. Compare each finger with the live hand model, then check the mapped robot joints within the selected operating envelope.

  • Review wearer fit and sensor placement.
  • Check the meaning of each angle, pressure channel and reference frame.
  • Repeat the same motion and look for slippage or inconsistent mapping.
Checking the glove input against a live hand model

Keep motion and contact together

Use a short sequence with visible events: approach, first contact, hold and release. Keep input timestamps, target commands, measured signals and video references in the recording so an unexpected grasp can be traced.

  1. Approach
  2. Contact
  3. Hold
  4. Release

Use research to frame the next experiment

Reis Carneiro and colleagues describe a tailored textile glove using printed strain and pressure sensors for robot interaction. Cerro and colleagues study a sensory glove that combines acoustic and motion signals with machine learning for assembly checks. These examples suggest useful questions about fit, signal choice and task events.

  • Which signal changes at the event you want to capture?
  • Does garment fit change the repeatability of that signal?
  • Can you review the raw input alongside the task result?
Discuss your application ↗