Decide

The Decide block serves as the culmination of the performance safety assurance, where all outputs generated through coverage and test evaluate are integrated into a clear and final decision: whether the system can be considered safe enough for approval and deployment. This decision is expressed as a binary outcome—pass or fail—regardless of the complexity and variability of preceding evaluation metrics. While earlier Safety Assurance Framework (SAF) blocks may involve graded performance indicators, confidence intervals, or probabilistic risk assessments, these must ultimately support a definitive, traceable, and auditable conclusion that aligns with regulatory expectations.
This approach is reflected in both regulatory and consumer-focused safety processes. In the case of ALKS (UN Regulation No. 157), the pass/fail decision is governed by explicit criteria—for instance, the vehicle must maintain a minimum time headway of 0.6 seconds and respond correctly to cut-in vehicles or emergency stops. Each test case has clearly defined thresholds, and failure to meet any criterion results in a binary fail outcome. On the other hand, Euro NCAP illustrates how a structured aggregation of many binary sub-decisions can feed into an overall performance score. For example, a failure to detect and brake for a pedestrian in a night-time scenario is a binary fail within the Autonomous Emergency Braking (AEB) assessment, which contributes to the final star rating. In both cases, while the presentation of results may differ—approval vs. rating—the underlying logic is consistent: multiple test outcomes are assessed against pass/fail thresholds, and the accumulated results inform a final binary outcome on system.
SAF Application Guidelines for 'Decide'
In the list below, “D” stands for Deliverable. All deliverables of the SUNRISE project can be found here.
- Verify binary decision output (D2.3 Section 4.3.4):
- Confirm that a clear pass/fail decision has been made about the CCAM system’s safety for approval and deployment.
- Ensure the conclusion is definitive, traceable, and auditable.
- Audit residual risk assessment (D2.3 Section 4.3.4):
- Check that residual risk has been properly identified.
- Verify the gap between all possible real-world scenarios (the “yellow space” in D2.3 Figure 11) and tested scenarios (the “red space” in D2.3 Figure 11) has been considered in the decision.
- Confirm decision logic integration (D2.3 Section 4.3.4):
- Ensure all outputs from the Coverage and Test Evaluate blocks have been properly integrated into the decision.
- Verify that any graded performance indicators, confidence intervals, or probabilistic risk assessments ultimately support the binary decision.
- Review acceptance of safety case arguments (D2.3 Section 4.3.4):
- Confirm that the Safety Case arguments justifying the gap between scenarios and test scenarios, have been properly evaluated.
- Verify that the acceptance of the Safety Case arguments has been properly documented.
