Develop and test drift-aware cadence checks

Develop and test a drift-aware ON/OFF cadence comparison or a detector that learns genuine cadence structure to distinguish terrestrial interference from genuine drifting signals without removing genuine detections at a comparable rate.

Background

The tested post-hoc cadence filter averages each ON and OFF panel over time before comparing them. This dilutes drifting signals and produces an ON/OFF statistic that poorly separates genuine detections from high-drift interference. The filter consequently removes genuine detections at approximately the same rate as false alarms.

The authors identify two concrete alternatives: integrate along the signal’s drift before comparing ON and OFF, as in de-Doppler pipelines, or train a detector to recognize genuine cadence patterns. Neither alternative was tested, leaving unresolved whether a cadence-aware method can retain genuine signals while rejecting interference.