Skip to content
TOBI
Skip to main content
Prototype / planned mass production information is subject to order confirmation

TOBI WRIST CAMERA / END-EFFECTOR VIEW

Wrist camera · End-effector viewpoint capture

The wrist camera supplements the first-person capture system with an end-effector view close to the hand and tools. It forms an integrable data node composed of monocular wide-angle video, IMU, camera parameters, and audio, suitable for detailed recording of manipulation processes such as grasping, placing, tool use, and dexterous hands. It can be combined with E6/E2 or used for standalone capture.

1280×1024Monocular video specifications
approximately 100gFull-unit weight specifications
200°Diagonal FOV specification
<1msPeripheral synchronization documentation
TOBI Wrist Camera Product Illustration (Official Diagram)
30 / 60fpsOptional video frame rate
174°Horizontal FOV specification
1kHzBuilt-in IMU (specs per version documentation)
SD cardLocal storage
<1msPeripheral synchronization documentation

PRODUCT OVERVIEW & CORE FUNCTIONS

Viewing the task from a head-mounted perspective,
View the action from a wrist-mounted perspective.

The wrist camera does not replace the headset; it supplements end-effector manipulation with visibility closer to the hand, gripper, and tools. It can capture independently or be used in combination with the E6/E2; the combination method, clock input, and peripheral connections must be confirmed for each specific project.

OPERATION-FOCUSED DATA

Video · IMU
Calibration · Audio

The documentation records support for color video, IMU, camera intrinsic and extrinsic parameters, and audio; infrared positioning lights and peripheral extensions such as gripper and glove tracking can be supported, creating fine-grained end-effector operation data nodes.

TOBI Wrist Camera (Official Schematic)
01 / Imaging

Monocular Wide-Angle View

Documentation specs are 1280×1024, 30/60fps; vertical 148°, horizontal 174°, diagonal 200° FOV, covering the near-range manipulation workspace.

02 / Depth of field

As close as 7.2 cm

Documented depth of field is 7.2 cm to infinity. Actual clear range is affected by lens, mounting, lighting, and target material.

03 / Data Nodes

Video and inertial synchronization records

Built-in IMU (metrics subject to version documentation); supports MP4/H.265 video, IMU, parameter, and audio data.

04 / Open expansion

External perception integration

Supports expansion with 2.4G, BLE, infrared positioning lights, and external perception devices such as grippers/data gloves; interfaces, protocols, and firmware versions are confirmed per project.

USE SCENARIOS

When "you can see the hand"
More important than "seeing the whole scene."

The wrist camera is suited to scenarios where fine end-effector motion details determine task success or failure. Whether to capture independently, whether to combine with E6/E2, and whether infrared positioning lights and gripper extensions are needed will be confirmed item by item during the solution phase.

Scenario 01

Grasp and place actions

Close-range recording of fingertip, gripper, and material contact, filling blind spots occluded by hands in the head-mounted view, preserving key action context for imitation learning.

Scenario 02

Dexterous hand / glove tracking

In combination with external perception devices such as glove tracking, hand poses and tool interactions during dexterous manipulation can be recorded; synchronization accuracy is documented as less than 1 ms, confirmed per combined configuration.

Scenario 03

Tool use and fine manipulation

For precision tasks such as assembly, tightening, plugging, and quality inspection, capture operational footage and audio near the tool end.

Scene 04

Combined Capture with E6 / E2

The head-mounted device provides global motion, while the wrist camera provides end-effector detail; once connected to host time synchronization, both data streams can be aligned on a single time base.

WHO IT SERVES

For teams that need to "see what the hands did."

The wrist camera is suitable for teams whose primary research or production focus is end-effector manipulation data, and it typically combines with a head-mounted capture device to form a complete solution. Whether it fits the task can be determined in a single combined technical evaluation.

Embodied AI / dexterous manipulation teams

Teams researching grasping, plugging/unplugging, and tool use need to incorporate end-effector vision and hand motion into training data; typically combined with E6/E2 head-mounted devices.

End-effector view · Composable

Data Capture and Evaluation Project Team

For collection projects that need to record both global and local manipulation details and require an open, verifiable data format, the wrist camera can be configured as a standardized data node in bulk.

All interfaces open · Batch

University laboratories and research institutions

Labs researching human-computer interaction, grippers and soft manipulation need a lightweight, scalable end-effector data acquisition unit that supports external perception integration.

Customizable · Peripheral expansion

WHY WRIST CAMERA & DATA BOUNDARY

Take the prototype and mass production and
State the combined capabilities clearly.

Documentation records that the wrist camera provided a prototype in July 2026, with mass production planned for August. The page discloses planned status, combination capabilities, and unverified capabilities separately, to avoid describing a single peripheral node as a complete standalone perception system.

WEIGHT

Lightweight wearability

100g

The overall device weight is approximately 100 g, so as a wrist node it imposes a low long-term wearing burden and does not interfere with end-effector manipulation.

FOV

Wide-angle coverage

200°

Diagonal FOV 200° (vertical 148° / horizontal 174°), with a depth of field down to 7.2 cm covering grasping and tool-manipulation ranges.

SYNC

Main unit combination synchronization

<1ms

When connected to E6/E2, outputs host synchronization information (PPS+UART TX); when used standalone, supports PPS+UART RX/TX. Peripheral synchronization accuracy is documented as less than 1 ms.

EXTEND

Exteroceptive extension

2.4G

Supports 2.4G, BLE, and infrared positioning lights; can be extended to connect external perception devices such as grippers and glove tracking.

Configuration note:If the goal is operational data for a dexterous hand or gripper, also confirm the wrist mounting position, lens coverage, time base relative to the head-worn device, external perception accessories, power supply, storage, and actual task samples. Prototype/production units, firmware, accessories, and delivery are subject to the order confirmation document.
Processor
4-core chip.
Camera
Monocular 1280×1024 at 30/60 fps; vertical FOV 148°, horizontal FOV 174°, diagonal FOV 200°.
Depth of field
Documentation records 7.2 cm to infinity.
Storage and data
Local SD card; color video (MP4/H.265), IMU, camera intrinsic and extrinsic parameters, and audio.
IMU
High-precision IMU (specifications subject to version documentation).
Wireless and interfaces
2.4G, BLE; supports infrared positioning lights. PPS and UART clock connection methods for host / standalone use are confirmed per the combination plan.
Synchronization Method
When connecting to E6/E2 for time synchronization, the wrist camera outputs host synchronization information (PPS + UART TX); when used independently, it supports PPS + UART RX/TX send and receive; peripheral synchronization accuracy is documented as <1ms.
Weight and System
Full device weight 100 g; Linux OS; interfaces, protocols, and customization scope can be confirmed per project.
Peripheral expansion
Supports integration with external perception devices such as grippers and glove tracking; module compatibility, power supply, synchronization, and software support are confirmed per project.
Status note: the documentation records prototype and planned mass-production timing; SD card capacity, battery life, wireless protocol details, included accessories, and final firmware version are not explicitly stated and are not default commitments on this page.

READY TO SPECIFY

Let the end-effector operation viewpoint,
Become a data node that can be included in the project acceptance definition.

The technical evaluation will jointly confirm the solution around task actions, wrist mounting, host combination, peripheral synchronization, data formats, and sample collection. This process does not constitute an automatic quote, supply commitment, or data performance guarantee.

Specify whether the wrist camera is used standalone or in combination with E6/E2.
Define the end task, mounting position, lens coverage, and sample scenarios.
Specify external gripper, glove, infrared light ring, and other extension requirements.
Clarify prototype/mass-production status, storage, firmware, accessories, and acceptance documentation.

DATA & PROJECT READINESS

From specifications to an acceptable capture plan.From specifications to a testable capture setup.

Can it be used directly for robot training?Is this data ready for robot training?

Device observations do not automatically equal robot action labels. First confirm the data fields, timestamps, coordinate frames, target action space and data quality, then validate the conversion and training pipeline.Device observations are not automatically robot-action labels. Validate fields, timestamps, coordinate frames, target action space and quality before conversion and training.

How do I confirm supply and delivery scope?How do I confirm supply and scope?

Provide the task, quantity and planned delivery date. Confirm the device version, firmware, accessories, support scope and acceptance method in writing. Historical prototypes or production plans do not mean current stock is available.Provide the task, quantity and delivery window. Confirm the version, firmware, accessories, support and acceptance method in writing. Historical production plans do not confirm current stock.

How do I sync with other devices?How is cross-device synchronization verified?

Starting together does not mean clock synchronization. For a combined solution, confirm the clock source, interfaces, timestamp definitions, error statistics and dropped-frame handling; single-device metrics cannot substitute for full-system acceptance.Starting together is not clock synchronization. Confirm clock source, interfaces, timestamp definitions, error statistics and dropped-frame handling for the complete system.