Planning a UWB real-time location system (RTLS) comes down to five decisions: how many base stations (UWB anchors) you need and where they go, how they are powered and networked, how their positions are calibrated, how the result is validated, and how often each tag reports.
Most problems found after go-live trace back to one of these decisions. This guide walks through each one with rules of thumb that apply to most UWB systems.
What do you need before you start?
A good plan starts with information, not hardware. Collect:
- A scaled floor plan (CAD or PDF) for every floor, with the tracked area and any zones with different needs.
- Accuracy and update-rate needs per zone, written as testable figures.
- Building details: ceiling heights, beams and cable trays for mounting, rack heights, large machines, doors, mezzanines and outdoor areas.
- What will be tracked: numbers of people, vehicles and assets, where tags will sit and how fast things move.
- Network facts: switch locations, free PoE ports and power budget, and how devices may reach the positioning server or cloud.
- Access constraints: lifts, work at height, production schedules and site rules.
If people will wear tags, add the privacy groundwork: a written purpose, employee consultation and a data protection impact assessment.
How many base stations do you need?
A simple formula gives a first estimate for indoor sites:
Base stations = the larger of 10 and (area in m² ÷ 200 × environment multiplier), rounded up
| Environment | Multiplier |
|---|---|
| Warehouse or distribution centre | 1.0 |
| Factory with metal structures | 1.2 |
| Factory with heavy machinery | 1.68 |
The minimum of 10 keeps enough base stations around even a small area for good geometry and some redundancy. Three worked examples:
| Site | Area | Multiplier | Estimate |
|---|---|---|---|
| Distribution centre | 2,000 m² | 1.0 | 10 base stations |
| Assembly plant with steel structures | 10,000 m² | 1.2 | 60 base stations |
| Heavy-machinery plant | 35,000 m² | 1.68 | 294 base stations |
Treat the result as a budget figure, not a layout. Open areas with line of sight can need fewer; long narrow aisles, mezzanines and several floors can need more. The base station estimator runs the same calculation for your area.
Where should base stations be mounted?
Geometry decides accuracy, so place base stations for geometry first and convenience second:
- Surround the area. Use the perimeter plus interior points, so that every spot sees base stations in several directions.
- Avoid straight lines. A single row along one wall or down the middle of a hall gives poor accuracy across it. Stagger base stations instead.
- Follow the aisles. In racked areas, make sure each aisle has line of sight to several base stations along its length, for example at both ends.
- Vary heights for 3D. If height matters, mount some base stations noticeably higher or lower than others.
- Choose the mounting point. Keep antennas clear of large metal surfaces, lamps and cable trays.
Mount high enough to see over people, vehicles and most stored goods, but below the level where full racks, crane runways or ducts block the view. On a square grid, one base station per 200 m² means about 14 m between neighbours.
How are base stations powered and connected?
Most fixed installations use Power over Ethernet: one cable from a PoE switch carries both data and power, so no mains socket is needed at each mounting point. Ethernet runs are limited to 100 m from the switch, which decides where switches go. Wire base stations in a star from the switches and check each switch’s total PoE budget; a UWB base station typically draws a few watts.
Agree the network design with IT early: a separate VLAN for the RTLS, the outbound connections the base stations need to reach the positioning server or cloud, and who manages the switches. Temporary, outdoor and sports set-ups can run on batteries or mains power with Wi-Fi or LTE instead of cabling.
How is the system calibrated?
The positioning engine must know where each base station is, to within a few centimetres, in the coordinates of the floor plan. Any error in those positions shifts every position computed from them.
| Method | How it works | Effort | Watch out for |
|---|---|---|---|
| Manual survey | Each base station is measured with a laser distance meter, total station or laser scanner | Grows with every base station; days on a large site | Measurement and typing errors; repeat after any move |
| Automatic calibration | Base stations range to each other and solve their own positions; a few measured points tie the result to the floor plan | Hours; little extra per base station | Base stations must see each other; still verify with reference points |
How do you validate the installation?
A validation walk proves the system against the requirement before anyone relies on it:
- Mark reference points with surveyed coordinates in each zone, including corners, aisles and doorways.
- Carry a tag at its real mounting height, a badge on the chest or a tag on a forklift roof, and pause at each point before walking the main routes.
- Compare reported and true positions, and calculate the median and 95th-percentile error per zone.
- Check events as well: zone entries and exits, alerts and head counts.
- Repeat during normal operation, with people, vehicles and full racks.
ISO/IEC 18305 describes metrics and test scenarios for this kind of evaluation. Fix gaps by adding or moving base stations before go-live, not after.
How do update rate and battery life trade off?
Every position costs the tag a few radio transmissions, so the update rate drives both battery life and how many tags an area can serve. Doubling the rate roughly doubles the radio energy a tag uses and halves the number of tags an area can serve at that rate. Tags with motion sensors can report less often, or sleep, when they are not moving.
| Update rate | Typical use | Trade-off |
|---|---|---|
| 0.1 Hz (every 10 s) | Parked assets, containers, inventory | Longest battery life |
| 1 Hz | People, pallets, carts, work in progress | A common default |
| 2–10 Hz | Forklifts and other vehicles, safety zones | Less lag, more energy |
| 10–50 Hz | Sports and fast machines | Shortest battery life; plan capacity |
How long does a deployment take?
| Phase | Typical duration |
|---|---|
| Planning and site survey | 1–2 weeks |
| Cabling and mounting | Days to several weeks, depending on area and access |
| Calibration | Hours with automatic calibration; days for a manual survey of a large site |
| Validation and tuning | 1–3 days |
| Integration and training | 2–6 weeks, in parallel with the above |
For a single hall, a few weeks from planning to use is realistic. Large and multi-building sites are usually rolled out in phases.
Where to go next
Before you place base stations, write down what accuracy each zone needs with the help of UWB accuracy and range, and get a first count from the base station estimator. To turn the plan into a budget, read RTLS cost and ROI. The hardware is described under base stations, and you can ask for a layout proposal for your floor plan.
Frequently asked questions
How many UWB anchors do I need?
How high should UWB base stations be mounted?
Can UWB base stations be placed in a line?
Do UWB base stations need network cables?
What is RTLS calibration?
How do you test an RTLS after installation?
How long does an RTLS installation take?
Sources and further reading
- R. B. Langley, Dilution of Precision, GPS World, May 1999
- ISO/IEC 18305:2016 Real time locating systems: test and evaluation of localization and tracking systems
- 47 CFR § 15.517: technical requirements for indoor UWB systems (US)
- Commission Implementing Decision (EU) 2019/785 on radio spectrum for ultra-wideband equipment
Published . Last reviewed by the Noccela engineering team.