Urban Data Platform × Edge AI Integration
On-site readings of urban change loop back as urban services.
Core infrastructure for the hyper-connected smart city — combining urban data, AR, and on-device AI.
On-site readings of urban change loop back into urban services. Smart-glasses and mobile sensors collect pedestrian-view data; on-device first-pass analysis links into the urban data hub; results return as AR guidance and operations dashboards.
— Technical Intent
On-site readings of urban change return as urban services.
The core is understanding the city through mobile edge devices. Smart-glasses collect pedestrian-view street and building data and analyze on-device.
— Architecture · 6 layers
Architecture
- 01 · Device
Mobile Edge Device
Smart-glass with camera, LiDAR, and IMU combined.
- 02 · Inference
On-Device AI
Run object detection, position estimation, event summarization on-site.
- 03 · Transfer
5G / Wi-Fi / LEO Path
Ground network by default, satellite as fallback.
- 04 · Broker
NGSI-LD Context Broker
Translate edge events into Linked Data.
- 05 · Hub
Urban Data Hub
Integrate fixed IoT, public data, and mobile sensor data.
- 06 · Service
AR City Service
Extend to AR HUD, city guidance, and safety info.
— Flow · 5 steps
Operating flow
- Step 01
Field Sensing
Smart-glasses collect location, gaze, and object data.
- Step 02
Edge Summarization
On-device AI summarizes raw data into events.
- Step 03
Context Update
The Context Broker converts events into NGSI-LD.
- Step 04
City Analytics
The Urban Data Hub combines fixed and mobile data.
- Step 05
AR Feedback
Analytics feed back as AR guidance and safety alerts.
— Operating Principles
Validation reliability for hyper-connected smart cities
Latency-Aware Edge
On-site decisions on-device; servers handle verification.
Privacy by Design
Minimize video transfer; anonymize; device authentication baked in.
Standards Integration
NGSI-LD ensures interoperability with the urban data hub.
Living Lab Service
Extends to living-lab services for tourism, transport, safety, welfare.
Other R&D
Back to CRAX →- R&D · 01
Multimodal Story-verse Creation Platform
Creation as a verifiable collaboration structure, not a single generation.
- R&D · 02
Conversational Multimodal AI for Media Production
Conversation becomes a production brief; the brief becomes an editable timeline.
- R&D · 03
Ontology-based AI Operations & Knowledge QA Platform
A structure where the basis of every answer is preserved.