Sensor Health and Calibration Monitoring
Objective
Detect a single sensor's own degradation or miscalibration over time, independent of multi-sensor disagreement.
Control / requirement
Continuous self-diagnostic/calibration-drift monitoring per sensor, with a defined recalibration or exclusion threshold.
Applicability
All physical AI systems with sensors subject to drift, wear, or environmental degradation.
Expected evidence
Calibration drift logs; recalibration/exclusion event records. [T2]
Assurance expectation
Evidence that drift thresholds are set below the point at which degraded input would affect safety, not only below sensor manufacturer tolerance.
Dependencies
PAI-SF-SAF-002 (model-drift-triggered safe-state) — sensor drift and model drift are related but distinct failure sources.
Exclusions
Not applicable to sensors with no meaningful drift/degradation mode (e.g., simple binary contact sensors).
Maturity / conformance relevance
Expected at Operational and High-Assurance levels.
Ecosystem relationship
Provides sensor-degradation incident categories for UAIF™, distinct from SEN-001's spoofing/disagreement categories.