How Automatic Mileage Detection Actually Works

An auto-detection mileage tracker decides on its own when you start a drive, follows it, and decides when you have stopped. All without you tapping a button. The how-it-works under the hood matters because it determines the trade-off between accuracy and battery.

Sources: Apple, startMonitoringSignificantLocationChanges(); Apple, CMMotionActivityManager; IRS Publication 463 for mileage substantiation requirements.

Step 1: Wake on motion

The app does not run continuously. Instead, it registers with iOS to be woken up when meaningful motion happens. Apple's SignificantLocationChange API fires when you move to a different "cell" of GPS space (typically 500m or more). CoreMotion can also wake the app when the phone detects vehicle-class motion.

Step 2: Confirm a trip is starting

Wake-up alone is not enough. You might be walking the dog or riding a bike. The app checks speed (Doppler-based, accurate even in a parking garage with poor GPS), motion type (CoreMotion classifies walking vs cycling vs driving), and rate of position change. Multiple signals confirm a vehicle is moving before recording starts.

Step 3: Record the route

During recording, the app pulls high-accuracy GPS positions (filtered for accuracy < 200 meters), runs them through a Kalman filter to smooth GPS jitter, and stores a sequence of route points. CoreMotion runs in parallel as a sanity check. If the phone goes still for too long, the app starts looking for a stop.

Step 4: Detect the stop

A combination of speed dropping to zero, position not changing for a configurable window (typically 2-3 minutes), and CoreMotion confirming non-vehicle motion ends the trip. A 120-second cooldown prevents accidental re-triggers when the driver pauses at a long red light or in a drive-through.

Step 5: Map matching at trip end

Raw GPS points usually do not align perfectly with road network. Map matching snaps the recorded points to the actual road network, producing a clean route polyline that matches the streets you drove. This is also when the final mileage figure is computed (route length, not GPS-point-by-GPS-point sum).

Why battery efficiency varies between apps

The biggest battery cost is continuous high-accuracy GPS. Apps that wake only on significant-location-change and escalate to high accuracy ONLY during confirmed driving can run all day with single-digit-percent battery drain. Apps that run continuous high-accuracy GPS will drain a phone in 6-10 hours.

Why accuracy varies

GPS accuracy degrades in urban canyons, parking garages, tunnels, and dense forest. Sophisticated trackers use Doppler speed (which works even when position is unreliable), dead reckoning during gaps, and the Kalman filter to maintain accuracy in low-GPS conditions. Less sophisticated trackers drop trips or record garbled routes when GPS quality dips.

The two Apple APIs this is built on

Auto-detection on iOS rests on two frameworks, and knowing which one an app leans on tells you most of what you need to know about its battery behaviour.

• Core Motion. CMMotionActivityManager reports what kind of motion the device is in, including an automotive state, using the motion coprocessor rather than the GPS radio. It is cheap to run continuously.
• Core Location. The significant-change location service wakes an app when the device has moved a meaningful distance, using cell and Wi-Fi positioning instead of continuous GPS. Apple documents it as the low-power option and can relaunch a terminated app to deliver an update.

An app that runs continuous high-accuracy GPS all day will detect trips reliably and drain the battery. An app that waits on motion and significant-change events costs far less power but has to reconstruct the first moments of a drive. That trade-off, not the app's marketing, is what determines both accuracy and battery cost.

What detection still cannot do for you

Detection produces distance, date, and route. It cannot produce business purpose, and IRS Publication 463 requires it. Every automatic tracker therefore needs one human input per trip, whether that is a classification swipe, a rule, or a category on a recurring destination. An app that claims to need nothing from you is either guessing at purpose or not recording it.

How accurate does the distance have to be?

The IRS wants a reasonable, substantiated figure rather than a certified one. Route-based distance from a GPS trace is comfortably defensible. Odometer readings at the start and end of the year support the total-miles element, which is why most well-built trackers ask for them even when they measure every trip.

See auto-detection done right. Start free with TruMile →

Related guides