Answers are grounded in this paper and cite their sources.
A design pattern for hosting AI evidence safely in safety-critical embedded perception: AI emits typed evidence only, a deterministic FSM holds exclusive publication authority, a CRC32 integrity gate degrades corrupted models to zero evidence, and every published transition carries an audit record. Validated by host replays of 45,566 frames of real UWB-radar captures (zero false confirmations; latch-up in all four weak-breathing sessions) and a simulated decision-level study showing conflict-driven abstention eliminates false alarms. Status: Preprint. This manuscript has been submitted to IEEE Embedded Systems Letters and is currently under peer review. It has not been accepted for publication. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this may no longer be accessible.
No takes yet. Share an insight, caveat, or question.
Lei Lv (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: