Live data from Hacker News

Ultra-low-power localization tag uses cellular signals [pdf]

cs.umd.edu

21–25 of 25 posts

Re: Ultra-low-power localization tag uses cellular signals [pdf]

#21
post #10
post #8

Earlier quoted context omitted.

Do you currently have tracker tags on those things? If not, why not? Multi-year lifetimes are already available with Tile (3 years), AirTag (>1 year), and Samsung's SmartTag 2 (700 days), using BLE detection instead of 4G detection. Millions of these would not degrade signal. A random bit of metal (like a quarter, as shown for scale) in between you an a 4G tower should have the same effect as one of these devices. Th…

Thanks. I miss-read it and thought it's energy harvesting from celular. I already have some trackers on important stuff (like keys and school bag of kids if they loose it in the bus). But for other use cases it's either too large (cats) or makes no sense to have too many because then I'd have to replace a battery every few weeks ("everything else").

I got a newish RF cat tracker lately and really like it: https://tabcat.com/

It's not app or Bluetooth based (has an included detector) and it works pretty well. Battery is supposed to last a few months. It easily fits on a cat collar and doesn't bother mine at all. I've been looking for something like this for years and someone finally made one!

Re: Ultra-low-power localization tag uses cellular signals [pdf]

#22
post #7

How are they able to achieve sub-100 meter accuracy? I might be missing something, but In my experience, collecting the cellular parameters is the easy part. The challenge is knowing where the cells actually are. You can ask Google's geolocation API, but that typically returns a fix with ~1000 meter accuracy if it returns a fix at all. These databases are curated based on GPS+cellular information from millions of pho…

Where I am (Australia) all the cellular base stations are licensed, and you can find the licence details including accurate lat/long online:

https://web.acma.gov.au/rrl/site_proximity.main_page

Re: Ultra-low-power localization tag uses cellular signals [pdf]

#23
post #13
post #12

Earlier quoted context omitted.

Interesting. CellMapper, at least, is not even a little bit open source. It doesn't even have an API. This leads me to wonder what other details that FA provides which might have been fabricated from thin air.

If someone is focusing on extracting data for a small area, they can manually browse the CellMapper data. An API isn’t required for that—it’s a matter of accessing and analyzing what’s already available through the platform.

Does that make it "open source"?

Re: Ultra-low-power localization tag uses cellular signals [pdf]

#24
post #23
post #13

Earlier quoted context omitted.

If someone is focusing on extracting data for a small area, they can manually browse the CellMapper data. An API isn’t required for that—it’s a matter of accessing and analyzing what’s already available through the platform.

Does that make it "open source"?

Kinda, yeah. It's open source in the "intelligence" sense of the phrase because it's openly available information (like OSINT), but it's not freely usable.

Re: Ultra-low-power localization tag uses cellular signals [pdf]

#25
post #14
post #9

Earlier quoted context omitted.

FTA: Our localization approach combines open-source databases (OpenCellID [47], Mozilla Location Services [66], and CellMapper [12]) for PCI-to-coordinate mapping of cell towers with a trilateration technique based on relative signal strengths.

That doesn't really answer the question. Those services basically work the same way as Google's geolocation API, but they're somehow able to magically get better accuracy.

Your assertion seems to be that Google's location service must be the best, and there's no way a trivial method can outperform it.

Have you considered that google might just perform poorly compared to other methods?

Post reply on HN