Why distance estimates move
Map accuracy in an indoor context is not a single number. It refers to how closely the digital representation of a space matches the physical reality, and whether that match is sufficient for the intended use. A floor plan imported from a building surveyor might be geometrically precise to within a few centimetres on paper, yet still fail to support reliable wayfinding if the coordinate system, scale, or alignment does not correspond to what the positioning engine actually produces.
Two distinct qualities matter. Geometric accuracy is the positional relationship between mapped features — walls, doors, columns, staircases — and their real-world locations. Functional accuracy is whether the map, when combined with the positioning system, correctly places a user in the right zone, triggers the right content, or guides them along a viable route. A map can be geometrically adequate but functionally poor if the positioning layer and the map are not calibrated to one another.
Validation is the process of testing that relationship under real conditions. It is separate from creating the map. You can produce a clean, well-structured indoor map and still find that users are consistently placed on the wrong side of a wall or that a zone trigger fires a metre too late. Validation catches those mismatches before visitors rely on the system.
A further distinction worth understanding is between absolute accuracy — where a point on the map corresponds to a real-world coordinate, such as a national grid reference — and relative accuracy — where the distances and angles between features on the map are correct in relation to each other, even if the entire map is slightly offset. Many indoor navigation systems rely on relative accuracy within a local coordinate frame, which is often sufficient for wayfinding but would not suit applications requiring survey-grade positioning.
Test the complete receiver path
Museums and galleries
In a museum, the question is usually whether a visitor standing in front of an exhibit is placed within the correct zone to receive the relevant audio or text content. The required accuracy depends on the spacing between exhibits. Where displays are several metres apart, a zone radius with a metre of tolerance may be acceptable. Where exhibits are closely packed, the margin for error shrinks considerably. Validation here means walking every exhibit position with a test device and confirming that the correct content triggers — and that content for a neighbouring exhibit does not.
Retail environments
Shop floor layouts change frequently. A map validated in April may be inaccurate by June if fixtures have moved. For zone-based notifications — such as triggering a promotion when a customer enters the footwear department — the critical check is whether the zone boundaries align with the physical department edges, not whether every shelf is plotted. Validation should focus on transition points: doorways, aisle ends and department boundaries where a user moves from one zone to another.
Events and temporary venues
At a conference or exhibition, the indoor map may be built from a layout plan rather than a surveyed floor plan. Validation timelines are compressed, and the infrastructure — staging, barriers, pop-up stands — may not be in place until hours before the event. Practical validation in this setting means identifying a small set of critical waypoints — entrances, registration, key stages, emergency exits — and confirming those first. Full aisle-by-aisle validation may not be possible, so the priority is ensuring that the routes visitors are most likely to need actually work.
Accessibility requirements
When indoor navigation is provided as an accessibility aid, accuracy takes on a different weight. A sighted visitor may tolerate being shown slightly the wrong position on a map and correct for it visually. A visitor relying on turn-by-turn audio instructions has fewer cues to fall back on. Validation for accessibility should include testing with the assistive technology and interface that end users will actually use — screen readers, haptic feedback, simplified map views — rather than only testing the standard visual interface.
Validation methods by context
The most common approach is a walking test: a tester follows predefined routes through the space with a device running the navigation software, recording where the system believes they are at each stage and comparing it to their known physical position. This is labour-intensive but reveals problems that static testing misses, because it captures the effect of movement, signal smoothing and handover between beacons or access points.
Static point testing is useful for establishing a baseline. The tester places a device at a known reference point — marked on the floor, measured from walls — and logs the reported position over a period, typically a minute or two. Repeating this across a grid of points gives a picture of accuracy and consistency across the space. The results should be recorded with the device model, orientation and any environmental conditions noted, because changing any of these variables can shift the results.
Use the result without overclaiming accuracy
Assuming the source plan is sufficient
Importing a CAD file or PDF floor plan and treating it as a validated map is a frequent error. Source plans may be out of date, drawn at a scale that introduces rounding errors, or use a coordinate system that does not align with the positioning engine. Even an accurate architectural plan may omit features that matter for navigation — temporary walls, furniture that blocks sight lines for visual positioning, or metal structures that affect Bluetooth signal propagation. The map must be validated in the environment it will serve, not merely against the drawing it was based on.
Testing with only one device
Different smartphone models have different Bluetooth radios, antenna positions and processing characteristics. A map that validates perfectly on one device may perform noticeably worse on another. Where the deployment will serve a public audience with varied devices, validation should include at least a small selection of common handset types. If that is not feasible, the validation report should note the device used and the limitation this introduces.
Confusing map accuracy with positioning accuracy
These are related but distinct. A perfectly accurate map paired with a positioning system that has a three-metre error radius will still place users in the wrong spot. Conversely, a positioning system performing within specification will produce misleading results if the map underneath it is misaligned. Validation needs to assess the combined system, not just one layer. If a user is consistently shown two metres north of their actual position, the cause could be the map offset, the positioning calibration, or both.
Ignoring environmental variability
Indoor spaces are not static. A venue validated on a Tuesday morning with minimal occupancy may behave differently on a Saturday afternoon when the space is crowded, doors are propped open, and additional stock or equipment has been introduced. For deployments where accuracy is critical, it is worth validating under at least two different occupancy conditions and noting any significant differences.
Not documenting validation results
Without a written record of what was tested, when, with what equipment, and what the results were, there is no basis for comparing future validations or identifying degradation over time. A validation record does not need to be elaborate. For each test point or route, note the location, the expected result, the actual result, the device used, and any relevant conditions. This creates a baseline that is invaluable when something stops working correctly months later.
Key checks to perform
- Alignment at known reference points: Stand at a clearly identifiable physical location — a doorway centre, a column, a marked junction — and check whether the map position matches.
- Zone boundary transitions: Walk slowly across each defined zone boundary and confirm that the system transitions at the expected physical point, not metres before or after.
- Route viability: Request routes between common origin-destination pairs and walk them, checking that the path does not lead through walls, closed doors or impassable obstacles.
- Consistency over time: At a single static point, observe whether the reported position drifts or jumps, which can indicate interference or instability in the positioning layer.
- Edge cases: Test near lifts, stairwells, glass partitions and large metal objects — the locations most likely to produce anomalous behaviour.
When to re-validate
Any physical change to the space — relocated fixtures, new partitions, changed signage — warrants at least a targeted re-validation of the affected area. Seasonal layout changes in retail, exhibition build-ups and breakdowns, and building maintenance work are common triggers. A full re-validation is also advisable after any update to the positioning infrastructure, such as replacing beacons, adjusting transmit power, or updating the positioning software, because changes in one layer can shift the calibration of the combined system.
