Three stacked layers. A physical edge of rugged sensors and edge processors; an intelligence core running real-time AI and an RTOS; and a digital cloud for analytics and model training. Data rises from the edge to the cloud, and trained models return to the edge, closing the loop.
- Layer 1 of 3: Physical edge. Rugged sensors and edge processors that survive heat, dust and vibration.
- Sensor array — Environment heat, dust, vibration
- Power stage
- Edge processor — Role acquisition and control
- I/O
- Actuator
- Layer 2 of 3: Intelligence core. A real-time AI + RTOS layer making zero-latency decisions at the edge.
- RTOS — Scheduling deterministic
- Inference — Decision point at the edge
- Control loop
- Layer 3 of 3: Digital cloud. Scalable analytics and model training that feed evolved models back to the edge.
- Analytics
- Model training — Output evolved models to the edge
- Fleet management
Connections:
- Sensor array delivers measurements to Edge processor.
- Edge processor issues actuation commands to I/O.
- Power stage delivers drive current to Actuator.
- Edge processor passes data up to RTOS.
- RTOS schedules Inference.
- Inference supplies setpoints to Control loop.
- Inference passes data up to Analytics.
- Analytics supplies datasets to Model training.
- Control loop passes data up to Fleet management.
- Model training returns trained models to Inference.
- Fleet management returns updated operating policy to Control loop.
