Comment from Persistence Analytics Group LLC
Persistence Analytics Group LLCSupportBusiness
Summary: Persistence Analytics Group LLC / United Grid supports NHTSA's objective of collecting safety-critical incident data for automated driving systems while seeking to reduce reporting burdens. The commenter argues that the reporting framework must prioritize high-quality, timely data that can identify when safety assumptions fail and ensure public transparency and data usability.
Comment on NHTSA-2026-0529
Incident Reporting for Automated Driving Systems and Level 2 Advanced Driver Assistance Systems
Persistence Analytics Group LLC / United Grid submits this comment regarding NHTSA’s information collection for incident reporting involving Automated Driving Systems and Level 2 Advanced Driver Assistance Systems.
PAG / United Grid supports NHTSA’s objective of collecting safety-critical incident data while reducing unnecessary reporting burden. Automated driving and driver-assistance technologies are moving quickly from controlled deployment into public operating environments. Incident reporting must therefore produce timely, usable, and decision-grade evidence.
The key issue is not only whether incidents are reported.
The key issue is whether the reporting framework allows NHTSA, state agencies, manufacturers, operators, insurers, researchers, and the public to identify when safety assumptions are failing.
NHTSA should evaluate the collection through an implementation-integrity standard:
1. Safety-critical data quality
Reports should capture enough information to determine whether the incident reflects system design, operational design domain limits, driver interaction, foreseeable misuse, environmental conditions, software performance, sensor limits, or deployment assumptions.
2. Timeliness and follow-up
Five-day reporting for higher-severity incidents is appropriate, but NHTSA should ensure that updated reports capture materially new information when crash facts, engagement status, vehicle damage, data availability, or narrative explanations change.
3. Engagement and handoff clarity
The 30-second engagement window is important because automated and human control may overlap during handoff, disengagement, mode confusion, or delayed driver response. Reports should help distinguish whether the safety issue arose before, during, or after automated-system engagement.
4. Vulnerable road users
Incidents involving pedestrians, cyclists, motorcyclists, wheelchair users, and other vulnerable road users should remain a high-priority reporting category. These incidents are especially important for public trust and deployment readiness.
5. Avoiding false confidence
Streamlining reporting should not eliminate information needed to detect recurring safety patterns. A lower-burden system is valuable only if it continues to reveal defect signals, misuse patterns, ODD limits, and emerging risks.
6. Data comparability
NHTSA states that the General Order is not primarily designed for apples-to-apples technology scorecards. Even so, the agency should continue improving public data presentation so incident trends, severity categories, system engagement, and follow-up actions can be understood without overstating what the data prove.
7. Public transparency and confidentiality
NHTSA should protect CBI and PII, but public reporting should remain useful enough to support safety oversight, state coordination, public trust, and independent review.
8. Verification after reporting
Incident reports should be treated as early-warning evidence, not final determinations. NHTSA should preserve the ability to request additional data, investigate recurring patterns, and identify when self-reported information is incomplete, inconsistent, or contradicted by other sources.
PAG / United Grid recommends that NHTSA maintain the reporting framework with continued attention to:
* safety-critical event capture;
* data-quality metrics;
* repeat-incident pattern detection;
* vulnerable-road-user protection;
* ODD and engagement-status clarity;
* updated-report completeness;
* public data usability;
* and follow-up verification when incident patterns suggest assumption failure.
The broader principle is simple:
Automated-vehicle safety cannot rely on market claims, deployment narratives, or self-assessed readiness alone.
It requires incident data that can reveal when technology, human factors, operating conditions, or deployment assumptions are not holding in the real world.
Modern mobility should advance.
But public-road deployment should be tied to decision-grade evidence, accountable reporting, and verified safety performance.
Respectfully submitted,
Neil P. Osnato
Founder
Persistence Analytics Group LLC | United Grid
National Security & Infrastructure Risk Analytics
Demand Durability | Grid Stress | Load Integrity
SAM.gov registered
UEI: D3VYU39H6DX9
D-U-N-S: 142849930
CAGE: 19T34
neil@persistenceanalyticsgroup.com
609-464-9055
https://persistenceanalyticsgroup.com/