Industries›Automotive
Automotive

Certify automotive AI before it reaches the vehicle

Driver-monitoring firmware and forward-perception models are scanned on every pull request, with the verdict right in GitHub.

The challenge

Perception and driver-monitoring models now ship in software updates, and every change must stay within ISO 26262 functional safety, SOTIF and the UNECE cybersecurity and software-update regulations.

Reviewing each release by hand does not keep up with the pace of pull requests.

Standards in scope

ISO 26262ISO 21448 (SOTIF)UNECE R155UNECE R156EU AI Act
How CertOne helps

Certification built into the way your teams already work

01

Checks on every pull request

A GitHub Action submits each change to CertOne. The verdict shows as a passing or failing check, and the findings appear in GitHub code scanning.

02

Firmware tested on its target architecture

Embedded ECU firmware runs under emulation of the ECU's arm64 processor, inside a hardware-attested confidential VM.

03

Findings mapped to the regulation

Each finding cites the clause it breaks, such as UNECE R155 or ISO 21448, and points to the file and line.

How your teams use it
Add the CertOne action

One step in the existing workflow.

Open a pull request

Each change is scanned in a confidential VM.

Read the verdict

In the checks and in code scanning.

Systems certified in the demo

Recorded on the live CertOne platform.

ADAS ECU Driver MonitorEdge

Embedded ECU firmware, a pure-C int8 driver-state model on arm64.

ADAS Forward PerceptionEdge

A forward-perception object detector (Tiny-YOLOv3, ONNX).

See CertOne on your industry's systems

Register to watch the full walkthroughs, recorded on the live platform.