Edge-to-Cloud AI for Connected Mobility
BLACKCYCLE
We're exhibitingMobility Playground, Hall 2 · 3–5 NovSmart city, from the junction up.
Deep tech on junctions already in service.
On-device inference
int8 on NPU. Latency measured on the board.
Real-time Linux
PREEMPT_RT, TSN, OP‑TEE.
HIL-validated control
Tested against a real traffic controller.
LLMs over live data
In production on MQTT telemetry.
Agents via MCP
Our own server. How we operate.
Training on digital twins
Multi-agent RL on SUMO. Dataset on Zenodo.
ETSI C-ITS stack
CAM, DENM, SPATEM, MAPEM, CPM.
C-ITS PKI, post-quantum ready
Roadmap published in IEEE Access.
pico-RSU
SkyV2XExisting pole. Controller untouched.
5 active lines · 2 published (IEEE Access, Zenodo)
Published research at Blackcycle Lab
What we put at a junction
One pole. Edge, radio and cloud.
Compact RSU
on the pole
Edge AI
OBSIA
over every junction
A junction already in service.
Pole, signal head, cabinet, controller. Capital the city has already spent.
One compact unit on the pole.
A PoE cable. Controller untouched. No civil works.
Signed messages to vehicles.
ITS-G5 or C-V2X. Every message signed on the unit.
Reads the signal state.
Edge AI on the pole. SPATEM, MAPEM, priority requests.
Cloud AI over every junction.
Live telemetry. Ask the junction why.
SkyV2X · pico-RSU
Upgrade city junctions with edge AI, cloud AI and V2X
Compact C-ITS roadside unit for junctions already in service.
- ITS-G5 or C-V2X · 5.9 GHz
- Messages signed on the unit
- Galileo OSNMA · RTK over V2X
- IP66 · PoE · 3.2 W
Prototype at Smart City Expo World Congress · Hall 2 · 3–5 Nov
Process
Analysis
Problem and objectives.
Design
Architecture.
Development
Implementation.
Delivery
Deployment and support.
Team
Miguel Fornell
Founder & CTO
Luis Carlos Morocho
Technical Business Developer
Carlos Saavedra
Strategic Advisor
Robinson Macias
Senior ML Engineer
Grace Mites
Technical Product Manager