GPS is a radio-navigation system designed around signals arriving from satellites high above the horizon. Buildings are therefore not just scenery; they change the radio environment that the receiver must solve.
Roofs, concrete, metal coatings and walls weaken satellite signals. A receiver may see fewer satellites or lower-quality measurements. Android may supplement that with Wi-Fi and cellular information, but the result is still not equivalent to an open-sky satellite fix.
The deeper you move inside a large building, underground station or parking garage, the less reasonable it is to expect a stable GPS trail.
A street lined by tall buildings resembles a canyon from the receiver's perspective. Some satellites are hidden, while other signals bounce from glass, steel or concrete before they reach the antenna.
Because GPS relies on extremely precise timing, an indirect path makes the signal appear to have travelled farther than it really did. The calculated position can shift toward the wrong side of the street or even into a building.
Imagine five successive fixes around your real position, each with a small error in a different direction. A map that connects those fixes draws a little polygon or zigzag. Nothing physically moved; the measurement error did.
Filtering can reduce visual noise, but aggressive filtering can also hide real short movements. Tracking software has to choose a balance appropriate to its purpose.
LocationOf displays the positions provided by the Android device. It cannot make satellite geometry obey a marketing promise. A strange indoor trail should be interpreted as uncertain sensor data, not as proof that a person walked through walls.
For general accuracy factors, see How accurate is phone GPS?.
When a map supplies an accuracy radius, imagine the phone as being somewhere inside that area rather than exactly at the center marker. The larger the uncertainty, the less meaningful it is to argue about a particular doorway, apartment or side of a street.
Route history can sometimes help more than the latest point. If several good outdoor fixes lead toward a station entrance and then the trail turns into a noisy cluster under the roof, the sensible interpretation is that satellite reception degraded after entering the building—not that the device repeatedly teleported around the concourse.
This matters when location is used for safety or operational decisions. Look for consistency across multiple timestamps, environmental context and the reported uncertainty. A tracker is a measurement tool, not an infallible witness.
If a map displays an accuracy radius, imagine the phone as being somewhere within that area rather than exactly at the centre marker. The larger the uncertainty, the less meaningful it becomes to argue about a particular doorway, apartment or side of a street.
Route context can be more informative than the final dot. If several stable outdoor fixes lead toward a railway station and then turn into a noisy cluster beneath the roof, the reasonable interpretation is that positioning quality deteriorated indoors. It does not mean the phone repeatedly jumped between every point in the cluster.
Filtering and sensor fusion can reduce obvious noise, but software cannot recover perfect measurements from radio signals that never reached the receiver cleanly. Aggressive filtering also creates its own trade-off: it can make a stationary trace look neat while smoothing away genuine short movements.
For important decisions, look at several timestamps, reported accuracy and the physical environment together. A phone tracker is a measurement tool with uncertainty, not an infallible witness.