E²-CAP-V
Multi-modal synchronized capture · millisecond timestamps · built for embodied-AI training
First-person capture built for embodied-AI training
Synchronized Stereo Capture
Wide-dynamic stereo cameras with synchronized dual 1920×1080 first-person capture — stable, detail-rich footage.
Record, Auto-save & Sync
30-second auto-segmentation (no lost clips even on power loss); auto-connects WiFi and uploads to the cloud once configured.
All-day Field Sessions
12000mAh power belt plus Bluetooth 5.2 pairing — built for long, multi-task on-site collection.
Complete kit, ready on arrival
Head-worn Capture Unit
Binocular camera worn on the head for first-person video.
Host (Main Unit)
Waist-mounted controller for recording, network setup and upload.
Power Belt
12000mAh battery that powers the host.
Cables
USB cable (head unit → host) + DC power cable (host → battery).
Designed for long-wear and on-site operations
Maximum Serviceability
External power support; replaceable straps and liners; one-click capture & upload via PC software; continuous firmware upgrades.
Ergonomic Design
Adjustable strap for an adaptive head fit; bio-inspired liner stays stable during rapid head turns.
Belt-mounted Power
Head unit and host stay light on the wearer; a 12000mAh power belt keeps capture going — ideal for multi-person, multi-rig sessions.
Technical Specifications
Capture Channels
Stereo wide-dynamic synchronized
Resolution
~3840×1080 (dual 1920×1080)
Frame Rate
~30fps
FOV
165°
Baseline
65mm (follows interpupillary distance)
Recording Format
MP4 (H.264)
Segmentation
30-second segments by default
Storage
SD card (exFAT), ~12–13 hrs on 128GB
WiFi
WiFi 5 (802.11ac)
Bluetooth
BT 5.2
Power
12000mAh belt supply
Operating Temp
−20°C ~ 70°C
Driving the embodied-AI data flywheel
Robot Skill Collection
Records human demonstrations to generate high-quality imitation-learning datasets, letting robots learn manipulation skills from real-world examples.
EGO Dataset Building
Large-scale first-person operation dataset production for VLA vision-language-action foundation models, building a prior knowledge base of human operations.
Behavior Research
Human operation analysis, gaze and attention modeling, and fine-motor rule studies — the scientific basis for robot cognitive modeling.
Turn first-person fieldwork into data assets
Contact us for E²-CAP-V capture solutions and pilot support.