RTK Experiments | Hacker News Reader

RTK Experiments

n1vux.github.io

33 pointsluu15 comments

I'm surprised Ardusimple isn't mentioned. It's heavily used in the opensource AgOpenGPS project.

Nice article and lovely piece of kit that one from sparkfun. Will likely be getting one next year to further my experiments with gunshot localization, on my Pi-based sound localization platform (https://github.com/hcfman/sbts-aru). Though at the distances I'm localizing the phone based GNSS seems to be working quite well.

With an RTk GNSS though I could do some experiments with localizing bat calls. I've tested the Pettersson Ultrasonic microphones with my localization platform and that works fine. I suspect that the bat localization I could achieve if my co-ordinates had RTK accuracy would be pretty damned good! But I have to try it to know.

I've been wondering if gunshot localization over Meshtastic would be worth pursuing. If you could just send timestamps and do all processing locally. Maybe you could send the RTK data via LoRa too? How bandwidth intensive is it?

Reminds me of a post I saw sometime ago: https://bendauphinee.com/writing/2023/12/30/to-know-where-yo...

>standalone GPS prices will be as staggering as real camera prices

Camera modules are dirt cheap, as are GPS modules. It's the screen and CPU to drive it that are expensive.

My unlocked Motorola Power G phone was only $200, which included both gps and a camera, along with enough CPU and battery to do everything I need. I assume there's no subsidy in that price.

It's development cost. Low volume hardware assembled in developed countries is always priced seemingly unreasonably due to worse amortization of engineering man-hours. Prices of products like low end phones built in developing countries from reused schematics nears costs of materials.

That gps is not RTK compatible.

(2022)

'Native GPS precision/resolution in Civilian mode is 30m; a 5 minute average can improve that to 5m'.

My iphone 15 mostly shows an error of 15 - 20 m inside and 8 - 12 m when I go outside. Never tested staying outside 5 min at one spot though. Still, it seems my current phone (late 2023) is more accurate than to be expected from this article?

There is a difference between accuracy and precision at play here, but modern gps is far more accurate than design intent.

It can also drop in accuracy all the way to 100m and often to 30m.

It can be very accurate (depending on conditions), but steady accuracy and momentary accuracy are two different things in practice.

Is it using WiFi, cell towers, and other location networks (like Glonass) for that? I suspect some of the difference might be cell phones taking any location signal they can get.

No, under 5m is the basic accuracy you get from an unassisted dual-band WAAS GPS on the ground, like in a standalone navigation device.

You have to remember that many of the error sources have been improved at the system level since GPS first launched:

1. WAAS and dual-frequency receivers correct the largest source of error, which is varying signal delay through the ionosphere. WAAS is broadcast on the same frequency as regular L1 signals, so receiving it and applying the corrections requires no hardware changes.

2. Satellite ephemeris and clock errors: New satellites have laser retro-reflectors so the control segment can measure their position more accurately. The control segment also uploads ephemerides more frequently, so they are fresher. Satellite clocks have been improved.

Once you correct for those, the main limitation is resolution in measuring time of arrival, which is where the high chip-rate military code helps. But carrier phase tracking can help here.

I don't know why commenters in this thread are quoting "30 meters". Maybe they got confused with the 3-sigma confidence in feet and then assumed meters as the units? You really have to be in an environment with serious multi path interference or GPS jammers or an extremely naive receiver implementation to get accuracy that bad.

It isn't just GPS, newer iPhones have hardware for all the main systems — GPS, GLONASS, Galileo and BeiDou.

The iPhone 15 Pro now has support for L5 GPS, and there's a (partially) operational constellation. This supposedly improves performance, but there's no real information beyond a mention on the iPhone stats page:

https://www.apple.com/iphone-15-pro/specs/

The "Precision dual-frequency GPS" is it, and the full constellation is scheduled to be operational in 2027.

Years ago GPS had indeed two modes of operation. However, also years ago, they announced they were dropping that civilian mode and everyone had the higher accuracy. Unless maybe there was another level.

Last year I saw a guy making measurements in a field. I asked him the price of his gtk unit. He replied 20,000 euros and it wasn’t even his best one. So 700 euros to get to play with this is really something.

Reply on news.ycombinator.com