Nordic Chipsets can support Indoor Location Tracking feature?

[Hemamali]

Hi Nordic Team, Can you please share any information regarding Indoor Location Tracking feature/application that nRF52840 can support? does this have AoA/AoD? and can this provide any kind of physical layer time-stamping for ToF/TDoA algorithms? What's the best location accuracy that can be achieved using this silicon? If i want to develop nRF52840 based Indoor Location Tracking application what kind of support i can get from you? What's the preferred module platform? does Nordic supports API for location developers? what's the best LOS & NLOS range one can get using this SoC? Kindly share this information

THanks & Best Regards, Hemamali

[endnode]

To BT5.0 and sudden improved NOLS: my believe is that inside complex topologies like office space the impact of long-range modulated air-speeds will be fairly minimal, but maybe I'm wrong. Also claim about AoA/AoD being achievable with "minimal HW deployment" I'm very skeptical because basically none of cheap low power BLE chipset have this today and just to get modulated air-speed it takes Nordic (one of the market leaders) up to 12 months to deliver production chip since specification announcement (note that nRF52840 is just in engineering samples' stage and it will hopefully available for mass market in the end of this year!). Anyway I'm big fan of these radio applications so I hope I will be wrong and all big hopes will come true.

[endnode]

To the examples: by 5 minutes of googling I found e.g. this blog where basically all the players are being interviewed: BluVision (now part of HID Global/ASSA ABLOY), Aruba Networks, Mist, Kontakt.io and others.

www.youtube.com/.../UCvcKPiivQ0KWz6vskAoAmdw

[Roger Clark]

You may want to consider the mathematics of the Time Of Flight calculations.

The speed of light is approx 300,000,000 meters per second in a vacuum and only slightly slower in air.

So for 1m resolution, you need a system clock of at least 300MHz and in real world scenarios, you would need at least 2 or 3 time higher clock rates, and probably more like 5+ time clock rate to get 1m resolution.

Plus you would need to have to synchronise the clock in both the TX and RX side, and hold those clocks stable to a reasonable accuracy over long enough periods of time ( unless you intend to constantly resync)

Don't get me wrong. This is technically possible with the correct hardware..

But general purpose Bluetooth MCUs are not the right hardware.

IMHO. To add the functionality you are looking for, to general purpose MCUs would hugely increase the cost and make them not commercially viable.

[endnode]

You are perfectly right and that's why chips with ToF are expensive systems. But AoA might be easier to achieve. And just to conclude: these companies claiming <2m RTLS accuracy are using various techniques but all are basically "lot of data processed heavily by the back-end" (seeing the same packet at different stations = different known locations and RSSI => trilat. + advanced filtering including knowledge of the topology from calibration phase => location).

[Roger Clark]

I've been looking at some unrelated RF devices, (motion detectors), and they have big problems with reflections when used indoors.

I suspect that AoA is also going to have the same sort of problem.

The walls in my house have a aluminium foil insulation layer in them, which is a great reflector ! But this is quite a common insulation technique, so I don't think I'm alone in noticing this.

There are some excellent dedicated hardware solutions giving decent ToF distance values, but they cost $$. I also doubt how useful they would be in normal indoor spaces which consist of multiple small rooms e.g. 10m x 10m.

So you'd need to deploy the hardware in each room.