We 7develop a location-tracking backend for sailing regatta tracking (RegattaHero, GPS tracking of racing sailboats via a Flutter iOS app). We rely on CLLocationManager for continuous position updates during races that can last several hours.
Problem: In light wind conditions, a sailboat can move at low but steady speed (e.g. 1-2 knots) with essentially no acceleration signature - unlike automotive or fitness activities, which involve intermittent acceleration/deceleration (starting, braking, footstrikes). We suspect Core Location's motion-based heuristics (built on accelerometer/motion coprocessor data, not solely GPS delta) interpret this near-constant, low-acceleration motion as "stationary" or low-priority, resulting in reduced update frequency or unexpected pausing of location updates.
We have already set pausesLocationUpdatesAutomatically = false explicitly, which per documentation should prevent automatic pausing entirely - but we still observe degraded update behavior in these conditions. This suggests the effect is not limited to the auto-pause mechanism but also influences other parts of the positioning/update-frequency pipeline that key off activityType.
We currently use activityType = .otherNavigation (the closest existing match, covering "boats" alongside cycling/trains/off-road vehicles), but this activity type appears tuned around vehicles with more typical acceleration profiles, not low-speed, low-acceleration marine drift.
Steps to Reproduce:
Set activityType = .otherNavigation, pausesLocationUpdatesAutomatically = false, desiredAccuracy = kCLLocationAccuracyBestForNavigation
Track a device moving continuously at 1-3 knots with minimal acceleration variance (e.g. sailboat drifting/sailing in light wind, or equivalent slow, smooth continuous motion) over an extended period (30+ minutes)
Compare update frequency/consistency to a scenario with typical stop-and-go motion (e.g. driving in traffic)
Expected:
Consistent location update delivery reflecting actual continuous movement, independent of acceleration variance, when pausesLocationUpdatesAutomatically is explicitly disabled.
Actual:
Update frequency/consistency appears to degrade during sustained low-acceleration, low-to-moderate-speed motion at sea, despite auto-pause being disabled.
Suggestion:
Either:
(a) Add a dedicated CLActivityType case for marine/nautical navigation, with pause/throttling heuristics based on speed-over-ground rather than acceleration signatures, or
(b) Document/expose a way to fully decouple update-frequency behavior from the motion-coprocessor-based heuristic when pausesLocationUpdatesAutomatically = false is set, so this flag reliably covers all related throttling behavior, not just the auto-pause feature itself.