UWB tracking: when sub-meter accuracy is worth the money, and when it is not
UWB is the technology you choose when a few meters of error would break the process, and when the system has to react in under a second. For those cases nothing else on this page substitutes for it. For the far more common case of finding equipment somewhere in a building, it is usually an expensive way to answer a question that 1 to 3 meters already answers.
UWB is not a system we developed: the products we build are the Bluetooth Beacon Tracker and the Sentinel Video Tracker. UWB we supply and integrate. Having no ultra wideband product of our own to defend is exactly why we can tell you when you do not need it.
The problems that genuinely need UWB
Four situations where meters of error, or seconds of delay, turn a working system into a useless one. If none of these describes your site, read the section on when UWB is a waste before you go any further.
A vehicle and a person can occupy the same space
Forklift anticollision is the clearest case. The system has to know where the truck and the pedestrian are, right now, and trigger in the time it takes a truck to travel a couple of meters. A position that is two meters off, or two seconds old, is not a safety function.
A zone boundary has to mean something
Restricted areas around presses, robot cells, overhead cranes or automated storage. If the boundary is enforced to within meters, the safe zone has to be drawn meters wider than it really is, which usually takes away the floor space the layout depended on.
The position drives a machine, not a person
Robotics, automated guided vehicles, tool positioning, assembly verification. When the consumer of the position data is software that acts without a human check, the tolerance is set by the process, not by what is convenient to install.
You need to know which station an item is at, not which area
Work in progress tracking on a dense production line, where stations are a couple of meters apart. If the technology cannot separate two adjacent stations, the data cannot be used to time the process.
Three questions to answer before you buy anything
These three separate the projects that need UWB from the much larger number that have simply been told they do. We work through them with you on the discovery call.
1. What happens if the position is 2 meters off?
If the answer is that a person walks a few steps further, you do not need sub-meter. If the answer is that a truck does not stop, or a robot cell arms itself late, you do. This is the question that decides the technology, and it takes one sentence to answer honestly.
2. Does anything have to react in under a second?
A search for a trolley can wait a few seconds. An interlock cannot. Our BLE system updates in the order of seconds, which rules it out of safety-critical control and rules it comfortably into everything else.
3. How large is the area that really needs this precision?
UWB cost scales with the area you cover, because it needs a dense network of anchors with careful geometry. Very often only one cell, one aisle or one doorway needs sub-meter, and the rest of the plant needs nothing of the sort.
How UWB positioning works, and what that implies
The mechanism explains both the accuracy and the cost. They come from the same place.
Note on figures: the section above describes how the technology behaves in general. We do not publish a UWB accuracy or update rate specification, because UWB is not a product we build and the real figures depend on the hardware chosen, the anchor layout and the site. Any number that reaches a proposal comes from the hardware we specify for you, and it gets validated in your pilot before anyone relies on it.
Time of flight, not signal strength
UWB measures how long a very short radio pulse takes to travel between a tag and an anchor, and converts that into distance. BLE systems infer distance from how strong a signal arrives, which is a far noisier quantity. This is why UWB is more precise, and it is a property of the method, not of the brand.
Anchors, and a lot of them
A tag has to be seen by several anchors at once, positioned so the geometry is not degenerate. In practice this means a dense grid, usually cabled for power and data, with a survey to fix the anchor coordinates. Coverage is not the design constraint, geometry is.
Active tags
A UWB tag transmits, so it has a battery, and the faster the update rate the shorter the battery life. That is a running cost per tracked item, and it is the reason UWB is rarely the answer for tens of thousands of low-value items.
Calibration is part of the install
The anchor positions have to be surveyed accurately, because the system can only be as good as the reference it was given. Sub-meter results are earned during commissioning. A proposal that treats installation as trivial is describing a different outcome from the one you will get.
When UWB is the wrong answer
We are not the ones who lose out if you buy UWB, and we still want to talk you out of it when the case does not hold. Four common patterns.
You are looking for equipment, not controlling a process
Infusion pumps, trolleys, tools, test instruments, containers. What the user does with the answer is walk to the right place and look. For that, 1 to 3 meters is enough, which is what our Bluetooth Beacon Tracker gives in 90 percent of cases, with beacons that run for years on their own battery. The gap between BLE and UWB here is a budget difference, not an operational one.
You need to cover a large or multi-building site
UWB cost follows area, so a full plant becomes an anchor and cabling project. The alternative for large coverage with little hardware is Wheere, our partner: according to Wheere, about 4 gateways cover up to around 1 square kilometer with sub-meter accuracy, using low-frequency radio at 148 to 174 MHz that passes through concrete. If the requirement is precision across a large site, that comparison is worth running before UWB.
Only knowing the room or the zone matters
If the operational question is which ward, which aisle, which department, then zone-level BLE or even a proximity mode will answer it at a fraction of the cost. Buying centimeters to answer a question asked in rooms is the most common way we see budget disappear.
You have thousands of items to tag
Active tags with batteries do not scale down in cost the way passive labels do. At high item counts the tag line dominates the budget, and passive UHF RFID at the chokepoints, or BLE beacons, will usually be the right economics. See our indoor asset tracking page for how that comparison plays out.
What UWB does not solve
Even in the cases where UWB is the right technology, there are things it will not do for you. These come up after installation, which is a bad time to discover them.
It does not survive a thin deployment
Accuracy depends on a tag being seen by enough well-placed anchors. Reduce the anchor count to fit a budget and the result is not slightly worse precision, it is unreliable precision in the gaps, which is harder to work with than a system that is honestly approximate everywhere.
It does not remove the need for a safety assessment
If the application is safety related, the positioning system is one component inside a safety case that also covers failure modes, response time end to end, what happens when a tag battery dies and what the machine does when data stops arriving. Buying accurate positions is not the same as delivering a safety function.
It does not fix an unowned process
This is true of every technology on this site. Precise data about items that nobody is responsible for returning, checking or acting on produces a precise record of the same problem.
It is not a budget technology stretched over a site
If the money available covers anchors for a quarter of the area you need, UWB is not the technology for this project. A BLE deployment covering all of it, or RFID at the points that matter, will serve the operation better than sub-meter accuracy in one corner.
What it costs
Our minimum commitment is USD 5,000. UWB projects normally sit well above that, and the reason is structural: the cost follows the area and the anchor density, not the number of things you want to track. Four factors decide it.
The area that needs sub-meter coverage
Anchors, mounting, power and data cabling for every one of them. This is the dominant line. Restricting UWB to the zone that genuinely needs it, and covering the rest with something cheaper, is the single largest saving available on most projects.
The site itself
Ceiling height, obstructions, metal structures and where power and network can actually be run. Two buildings of identical size can need very different anchor counts, which is what a site assessment is for.
Number of tags and update rate
Active tags have a recurring battery cost, and a faster update rate shortens battery life. The update rate should be set by the process requirement, not chosen at the top of the range by default.
Commissioning and integration
Anchor surveying, calibration and connecting the output to your systems. On UWB projects the commissioning effort is a larger share of the total than people expect, and it is what separates a system that reaches its specification from one that does not.
How we work
Four steps. On UWB inquiries the first step quite often ends with a different technology, and that is a normal outcome rather than a failed sale.
01. Free 30-minute discovery call
We work through the three questions with you, starting with what happens if the position is 2 meters off. If UWB is not justified, you will hear it here.
02. Site assessment
Floor plan in scale with the areas that need which level of precision, the constraints on mounting and cabling, and a few photos. This is where a mixed design usually appears: precision where it is needed, something cheaper everywhere else.
03. Pilot
A real deployment on a limited area, because vendor specifications are not site results. Where the pilot is BLE, that means at least 6 nodes covering roughly 300 square meters, which is enough to show you whether 1 to 3 meters is sufficient before you spend on anything more precise.
04. Rollout and integration
Full deployment and integration with your systems. On our own BLE system that means a REST API, push callbacks and optional daily CSV files, with output as WGS 84 coordinates or metric X and Y. The raw data is yours.
Frequently asked questions
How accurate is UWB in practice?
UWB is what organizations specify when the requirement is typically under one meter, and in a well-commissioned deployment that is what it is bought to deliver. We do not publish a figure of our own, because UWB is not a product we build and the real result depends on the hardware, on anchor geometry and on the building. Any number in a proposal should arrive with a pilot that confirms it on your site. We measure it against your process in your building, rather than quoting a number from someone else's lab, and that applies to our own products too.
Do you supply UWB hardware?
Yes. Our own products are BLE and video, so ultra wideband is not something we built, but it is something we supply: we assess it, specify it against the alternatives, supply it and integrate it, and we run the project end to end. Not having built it is the reason we can tell you when you do not need it: there is no ultra wideband product of ours to protect, so the recommendation follows the requirement. What your site actually needs is settled on the discovery call and confirmed in the pilot.
Why is UWB more expensive than BLE?
Because the cost follows the area rather than the number of tags. UWB needs a dense grid of anchors with good geometry, usually cabled, plus a survey to fix their coordinates. BLE nodes sit on the ceiling, use Wi-Fi at 2.4 GHz or Ethernet with USB power, and can work at a much lower density because the requirement they serve is 1 to 3 meters rather than centimeters.
Can UWB cover an entire plant?
Technically yes, at a cost that scales with the floor area. Most sites are better served by a mixed design: UWB or another precise technology in the cells that need it, and something cheaper for the rest. Where high accuracy really is needed across a large area, the alternative worth comparing is Wheere. According to Wheere, about 4 gateways cover up to around 1 square kilometer with sub-meter accuracy and the signal passes through concrete. We are a Wheere partner and we will run that comparison honestly.
Does forklift anticollision require UWB?
It requires a system that reacts in under a second with precision in the order of a meter, which rules out our BLE system: it updates in the order of seconds and we will not sell it for that use. Whether UWB or a dedicated proximity detection product is the better answer depends on the site, the traffic and the safety case, and it is a decision that belongs with your safety assessment rather than with a positioning vendor.
What is the minimum budget?
USD 5,000 is our minimum commitment. A UWB deployment is normally well above it, which is one more reason to confirm on the discovery call that you need the precision before you price the hardware.
Specifications
Accuracy: 1 to 3 meters in 90 percent of cases
Update rate: Order of seconds, not sub-second
Tags: Standard iBeacon, years of battery life, range up to 50 meters
Nodes: Ceiling mounted, detect beacons up to 30 meters, Wi-Fi 2.4 GHz or Ethernet, USB 5V power, minus 20 to plus 60 degrees C
Ceiling height: Works up to 7 meters
Buildings: Up to 30 levels, one floor plan per level
Output: WGS 84 coordinates or metric X and Y
Integration: REST API, push callbacks, optional daily CSV with positions every 5 seconds
Hosting: Cloud by default, on-premise supported
Pilot: At least 6 nodes, covering roughly 300 square meters
Talk to an engineer, not a salesperson
Tell us what you need to locate, how precisely and how fast the system has to react. If UWB is the answer we will say so, and if 1 to 3 meters is enough we will say that instead.