Technology Trends
Indoor positioning technology works out where a person or object is inside a building, where GPS cannot. An indoor positioning system (IPS) uses signals available indoors — radio, magnetic, inertial — to place a device on a floor plan, delivering the blue dot that people expect from outdoor maps. Accuracy ranges from tens of centimetres to several metres depending on the technology and how much hardware you are prepared to install.
This guide is the hub for the topic: why GPS fails indoors, how indoor location tracking is actually calculated, how the technologies compare on accuracy, what drives the cost of a deployment, and the difference between an indoor map, indoor positioning and indoor intelligence — three things routinely confused during procurement. It does not quote prices: the drivers vary by an order of magnitude between buildings, and a published figure would mislead more than it helped.
Why GPS does not work indoors
GPS satellites transmit from roughly 20,000 km away, and by the time those signals reach the ground they are extremely weak. A roof, steel frame or concrete floor attenuates them below usable levels, and what does get through has usually bounced off surrounding surfaces — so the receiver calculates from a signal that took a longer path than the direct one.
The US government’s own gps.gov puts smartphone GPS at “typically accurate to within a 4.9 m (16 ft.) radius under open sky”, notes accuracy “worsens near buildings, bridges, and trees”, and lists satellite signal blockage, multipath reflection and “indoor or underground use” among its error sources. It publishes no indoor accuracy figure, because there is no specified indoor performance. Any article quoting one has invented it.
How indoor positioning is calculated
Whatever the signal, the maths falls into four families — and the choice matters more than the radio:
- Time of flight — measure how long a signal takes to travel, convert to distance. The most accurate approach, and why ultra-wideband leads: its very short pulses allow far more precise timestamping than narrowband radio.
- Angle of arrival — an antenna array measures the direction a signal came from; intersecting bearings give a position.
- Signal strength (RSSI) — infer distance from signal weakening. Cheap and universally available, but walls, people and metal attenuate unpredictably.
- Fingerprinting and sensor fusion — compare what the device currently observes (Wi-Fi, magnetic field, motion sensors) against a surveyed map of the building, and fuse it with the phone’s own accelerometer and gyroscope to smooth the result. This is how positioning can work without installing anything.
How accurate is indoor positioning?
| Approach | Typical accuracy | Source | What it costs you |
|---|---|---|---|
| UWB | 10–30 cm controlled; ~50 cm deployed | Zafari et al.; FiRa; Sensors 2020 | Dense anchor infrastructure |
| BLE / beacons | 0.7 m with direction finding; 1–3 m with RSSI | Pau et al., Sensors 2021; Cisco | Beacons to install, maintain, re-battery |
| Hardware-free (sensor fusion) | 1–5 m | Mapsted, measured across deployed sites | No in-building installation |
| Geomagnetic | 1–5 m | Mapsted, measured across deployed sites | Survey effort; affected by structural change |
| Wi-Fi | 1–2 m with 802.11mc RTT; 3–5 m with RSSI | Android Developers; Cisco | Reuses existing APs |
| RFID | Proximity to zone level; active range up to ~100 m | Zafari et al. | Readers at fixed points |
| 5G | Sub-metre in principle; highly deployment-dependent | See our 5G article | Requires in-building 5G |
| GPS (for comparison) | ~4.9 m open sky; no specified indoor performance | gps.gov | Outdoors only |
Note that the two rows sourced to Mapsted are our own measurements rather than independent ones — treat them as you would any vendor figure, and ask for a site survey. Two further warnings about accuracy numbers, including ours. First, laboratory accuracy is not deployed accuracy: the FiRa Consortium describes UWB “down to a few centimeters” in good conditions, while a peer-reviewed automotive deployment reported in Sensors measured 50 cm with eight anchors per 2,000–3,000 m². Both are honest; they measure different situations. Second, ask what anchor density and line-of-sight conditions a quoted figure assumed. For the full sourced comparison see our RTLS guide.
Indoor maps, indoor positioning and indoor intelligence
These three get used interchangeably in procurement and they are different things. Buying one when you needed another is a common and expensive error.
| What it is | What it answers | |
|---|---|---|
| Indoor map | A digital model of the building — levels, rooms, corridors, and how they connect | “What is in this building and how do the spaces relate?” |
| Indoor positioning | Determining where a device is on that map, live | “Where am I, and where is that thing?” |
| Indoor intelligence | Analysis of positioning data over time | “How is this building actually used, and what should change?” |
The dependency runs one way. A map without positioning is a static floor plan — useful for signage and space planning, but it cannot guide anyone. Positioning without a map produces coordinates nobody can interpret. And indoor intelligence is not a separate purchase: it is what the positioning data becomes once you retain and analyse it, which is why the retention and export questions belong in the original procurement rather than a later one.
What indoor intelligence actually tells you
- Dwell and flow — where people stop, how long for, and the routes they actually take rather than the ones the design intended
- Congestion and bottlenecks — where queues form, and when
- Space utilisation — which areas are busy, which are dead, and how that varies by hour and day
- Journey completion — where people abandon a route, which usually means they got lost there
For a retailer this informs layout and staffing; for a hospital it exposes where wayfinding is failing; for a workplace it answers how much space is genuinely needed. One caution worth stating: this data describes people, so anonymisation, aggregation and retention policy are part of the design, not an afterthought.
Where indoor positioning is used
Warehouses and distribution centres
Large, repetitive, metal-rich buildings — among the hardest environments for radio positioning, and among the most valuable. Typical uses are locating stock and equipment, routing pickers efficiently, forklift safety and collision avoidance, and measuring where travel time is actually spent. Racking and moving metal make this a place to be sceptical of lab accuracy figures and insist on a site survey.
Retail and venues
Product-level wayfinding, footfall and dwell analytics, and proximity-triggered offers. The analytics usually justify the deployment before any customer-facing feature does.
Healthcare
Patient and visitor navigation, staff duress, and equipment location. See hospital wayfinding and RTLS in healthcare.
Airports, campuses and offices
Passenger and visitor wayfinding across large multi-building estates, desk and room utilisation, and emergency mustering — where a live roll-call replaces a manual headcount.
How IoT fits in
“IoT indoor positioning” usually means one of two things. Either the positioned objects are IoT devices — tags on equipment reporting location alongside telemetry such as temperature, motion or battery state — or building IoT systems consume position data, so lighting, HVAC and access control respond to where people actually are.
The practical consequences are worth planning for: sensors beyond location shorten battery life, so specify only the telemetry you will act on; and anything joining the network belongs in your security review and segmentation plan alongside other connected equipment.
What does an indoor positioning system cost?
There is no honest per-square-metre price, because the drivers vary by an order of magnitude between deployments. What determines the number:
- Accuracy tier — the single biggest lever. Coordinate-level accuracy requires dense anchors; zone level often requires none.
- Floor area and building count, and how much of it needs coverage.
- Infrastructure — beacons or anchors to buy, install, power, calibrate and re-battery, plus the cost of doing that in an operating building.
- Mapping and survey — usually underestimated, and recurring, because buildings change.
- Integration with the app, CMMS, access control or analytics stack.
- Tags, if objects rather than phones are being located.
Ask every vendor to quote against your floor plans and to break out installation and ongoing maintenance separately from licence cost. Where the estate is large or cannot be closed for an installation programme, the infrastructure line is frequently larger than the software one.
Where Mapsted fits
Mapsted’s positioning is hardware-free: it locates a phone using sensor fusion rather than beacons, Wi-Fi dependency or anchor infrastructure installed through the building, at an accuracy of roughly 1–5 metres — room and zone level, not centimetres. For asset and personnel tracking there are Mapsted Tag and Mapsted Badge, and Mapsted Flow for anonymous movement analytics.
Being explicit about the trade-off: if you need dependable centimetre accuracy — instrument positioning, safety interlocks around moving plant — a dense UWB deployment is the architecture built for that and we will say so. For wayfinding, analytics and zone-level tracking, which is most requirements, avoiding an in-building installation is usually the larger practical win. Tell us about your building and ask us to prove it with a site survey rather than a claim on a page.
Frequently Asked Questions
What is indoor positioning technology?
Technology that determines the location of a person or object inside a building, where GPS is unreliable. It uses signals available indoors — radio, magnetic field, device motion sensors — to place a device on a digital floor plan in real time.
What is the difference between an indoor map and indoor positioning?
An indoor map is the model of the building — levels, rooms and how they connect. Indoor positioning determines where a device is on that map. A map alone cannot guide anyone; positioning alone produces coordinates with no context. Wayfinding needs both.
How accurate is an indoor positioning system?
From about 10–30 cm for UWB in controlled conditions to 3–15 m for basic Wi-Fi positioning. Hardware-free, BLE and geomagnetic approaches typically land at 1–5 m. Real deployments generally perform worse than laboratory figures, so ask what conditions a quoted number assumed.
Does indoor positioning need beacons?
Not necessarily. Beacon-based systems need hardware installed, powered and maintained throughout the building. Sensor-fusion approaches use the sensors already in the phone and need no in-building installation, at the cost of the very highest accuracy tier.
What is indoor intelligence?
The analysis of indoor positioning data over time — dwell, flow, congestion, utilisation and journey completion. It is not a separate product so much as what positioning data becomes when retained and analysed, which is why data retention and export should be settled in the original procurement.
Can indoor positioning work in a warehouse?
Yes, and it is a common application — stock location, picker routing, forklift safety and travel-time analysis. Warehouses are also among the hardest radio environments, because racking and moving metal cause reflection and blockage, so insist on a site survey rather than relying on a published accuracy figure.
