Building a local positioning system to track runners using Ultra-Wideband

We build asset tracking software solutions. We tend to work with a lot of third parties for hardware tracking solutions. I know Pozyx who are mentioned in the article, they are a good example of a company active in this space. Nice technology stack and competent people.

UWB as a product category is very broad. There are many use cases and different takes on how to do positioning. And some use cases go well beyond just positioning. UWB is actually being used for biometrics. There are people doing things like measuring heart rate with UWB.

For positioning there are broadly two approaches. Apple and Google are backing standards via the FIRA consortium. These are smart phone centric. The phone determines its distance to one or more UWB things (car keys, anchors, ticketing machines, luggage tags etc.) and then you do something with that information in an app. The big limitation here is privacy. There is almost no API ecosystem around tags because Apple and Google have completely locked down everything to prevent third parties from building things that could enable people tracking. That's why industrial usage of Apple tags is basically non existent even though there would be plenty of use cases; if only you could get your hands on some APIs.

And then there's the classic use case that is more anchor centric where triangulation happens by combining signals from multiple anchors with a known location. This system is a good example of that. Here the big trend is that WIFI base station makers are increasingly also bundling UWB anchor functionality. This lowers the cost of infrastructure needed in buildings to provide coverage. Tags are battery operated and typically motion activated, they usually have some kind of firmware, and in some cases support e.g. blinking leds that you can trigger via RPC calls. However, these systems tend to be very expensive to install and generally require bespoke software solutions. There's very little that works off the shelf that actually does something useful.

A big challenge from a software point of view is that SDKs are primitive, mostly proprietary, vendor specific, and heavily biased towards on premise networked solutions. IMHO this is holding back the whole ecosystem.

Standardization and more open technology stacks would help. But a bigger problem is that the world is messy and there is no single vendor that covers all of it. You need hybrid solutions.

The 24-hour run (counterpart of this event from University of Leuven) has their counting system open-sourced here (although it's no longer used I think):

https://github.com/ULYSSIS-KUL/ipp/

Great blog post @robinpdev!

> 10 USD per Module isn't really competitive with disposable RFID. Great for a relay though it seems!

This is a really cool idea, and now I'm thinking about adapting this for racecars; the telemetry and live tracking possibilities seem fantastic.

uwb is pretty neat. in the last few months i started getting 3/4ths of a gigabit of downstream bandwidth on a lot of nyc streets on my ordinary verizon plan.