Sensing, deciding and acting inside one timing budget.
Automotive systems fail at the seams between subsystems. The work is in the interfaces and the timing, not in any single component.
WHAT THE WORK TURNS ON
- Sensor fusion
- Multiple sensing streams reconciled into one view the controller can act on.
- ADAS interfaces
- Driver-assistance functions treated as a system interface with defined behaviour.
- Autonomous platforms
- Compute, sensing and power as one architecture with a shared thermal and timing budget.
ARCHITECTURE
The architecture this sector draws on. It is the same system TrustXP builds in the corresponding domain — the sector changes what matters in it, not what it is.
IoT and embedded system architecture
- Device
- Sensor array and the embedded controller that reads it.
- Edge intelligence
- An RTOS and an inference layer deciding on-device.
- Cloud services
- Integration, device management and over-the-air update.
A sensor array feeds an edge controller running an RTOS and an inference layer. Connectivity carries data to cloud services, and a device-management path returns firmware and configuration over the air.
- Layer 1 of 3: Device. Sensor array and the embedded controller that reads it.
- Sensor array — Concern deterministic control loops
- MCU / SoC — Concern embedded firmware
- Peripherals
- Layer 2 of 3: Edge intelligence. An RTOS and an inference layer deciding on-device.
- RTOS — Concern edge inference latency
- Inference — Validation medical-device controls
- Layer 3 of 3: Cloud services. Integration, device management and over-the-air update.
- Connectivity
- Device management — Path OTA update
Connections:
- Sensor array reports readings to MCU / SoC.
- MCU / SoC drives Peripherals.
- MCU / SoC passes data up to RTOS.
- RTOS schedules Inference.
- Inference passes data up to Connectivity.
- Connectivity carries telemetry to Device management.
- Device management returns firmware and configuration to Inference.