Garmin Speed sensor

On Thu, 11 Feb 2021 12:50:36 +0100, Rolf Mantel
wrote:

Am 11.02.2021 um 06:05 schrieb John B.:
I read that the top quality
professional GPS receivers will plot to within 7 inches of a true
location but will a "made for the masses" device like a hand phone or
a bicycle meter do as well?

This kind of accuracy is only possible via "differential GPS", not from
the standard GPS algorithm.

Generally true for consumer GPS receivers that are moving. 7 inch (18
cm) accuracy requires access to military encrypted frequencies and/or
additional post-processing:

Accuracy
(meters)
GPS only 20
WAAS corrections 3
GLONASS 3
GALILEO 1
BeiDou 3.6

However, for stationary accuracy, such as for surveying, it is
possible to reduce errors by averaging over time. I use a program on
my phone to do this:
https://play.google.com/store/apps/details?id=gr.stasta.mobiletopographer
Looking at my numbers for a 10 minute run at my house and using GPS,
WAAS, GLONASS and GALILEO, I'm getting about 1.4 meters horizontal
accuracy:
http://www.learnbydestroying.com/jeffl/crud/GPS-averaging.png
The phone is a Moto G Power (2020).

The practicality of using averaging to increase accuracy for tracking
users is rather dubious. The problem is that the phone needs to be
running the GPS and averaging software continuously to be effective.
These will rapidly deplete the battery, which should make the owner
rather suspicious.


--
Jeff Liebermann
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558

On Mon, 15 Feb 2021 12:28:25 -0800, Jeff Liebermann scribed:

On Thu, 11 Feb 2021 12:50:36 +0100, Rolf Mantel
wrote:

Am 11.02.2021 um 06:05 schrieb John B.:
I read that the top quality professional GPS receivers will plot to
within 7 inches of a true location but will a "made for the masses"
device like a hand phone or a bicycle meter do as well?

This kind of accuracy is only possible via "differential GPS", not from
the standard GPS algorithm.

Generally true for consumer GPS receivers that are moving. 7 inch (18
cm) accuracy requires access to military encrypted frequencies and/or
additional post-processing:

Accuracy (meters)
GPS only 20 WAAS corrections 3 GLONASS 3 GALILEO
1 BeiDou 3.6

However, for stationary accuracy, such as for surveying, it is possible
to reduce errors by averaging over time. I use a program on my phone to
do this:
https://play.google.com/store/apps/details?
id=gr.stasta.mobiletopographer
Looking at my numbers for a 10 minute run at my house and using GPS,
WAAS, GLONASS and GALILEO, I'm getting about 1.4 meters horizontal
accuracy:
http://www.learnbydestroying.com/jeffl/crud/GPS-averaging.png
The phone is a Moto G Power (2020).

The practicality of using averaging to increase accuracy for tracking
users is rather dubious. The problem is that the phone needs to be
running the GPS and averaging software continuously to be effective.
These will rapidly deplete the battery, which should make the owner
rather suspicious.

If you are using 'back processing' and can afford it, for the equipment
and subscription(? dated knowledge), you can achieve in the order of 1mm/
a few mm accuracy.

Surveyors realised very quickly that all they had to do is adjust the
received signal by the offset of a known LOCAL point.

Since the offset varies over time, Initially, you could download tables
of the offset for every 5(?) minutes of a particular day. Now, with the
correct equipment, you tune your equipment to the live back channel(sub
carrier on various AM/FM/?) broadcast to have 'instant' corrected data.

Averaging over time for a day/week/month, with a PC is good.
I've also manualy adjusted from the downloadable tables, but I've never
had the use to justify the cost of the live back channel gear.

Am 16.02.2021 um 00:29 schrieb News 2021:
On Mon, 15 Feb 2021 12:28:25 -0800, Jeff Liebermann scribed:

On Thu, 11 Feb 2021 12:50:36 +0100, Rolf Mantel
wrote:

Am 11.02.2021 um 06:05 schrieb John B.:
I read that the top quality professional GPS receivers will plot to
within 7 inches of a true location but will a "made for the masses"
device like a hand phone or a bicycle meter do as well?

This kind of accuracy is only possible via "differential GPS", not from
the standard GPS algorithm.

Generally true for consumer GPS receivers that are moving. 7 inch (18
cm) accuracy requires access to military encrypted frequencies and/or
additional post-processing:

Accuracy (meters)
GPS only 20 WAAS corrections 3 GLONASS 3 GALILEO
1 BeiDou 3.6

However, for stationary accuracy, such as for surveying, it is possible
to reduce errors by averaging over time. I use a program on my phone to
do this:
https://play.google.com/store/apps/details?
id=gr.stasta.mobiletopographer
Looking at my numbers for a 10 minute run at my house and using GPS,
WAAS, GLONASS and GALILEO, I'm getting about 1.4 meters horizontal
accuracy:
http://www.learnbydestroying.com/jeffl/crud/GPS-averaging.png
The phone is a Moto G Power (2020).

The practicality of using averaging to increase accuracy for tracking
users is rather dubious. The problem is that the phone needs to be
running the GPS and averaging software continuously to be effective.
These will rapidly deplete the battery, which should make the owner
rather suspicious.

If you are using 'back processing' and can afford it, for the equipment
and subscription(? dated knowledge), you can achieve in the order of 1mm/
a few mm accuracy.

Surveyors realised very quickly that all they had to do is adjust the
received signal by the offset of a known LOCAL point.

This is what is decribed in Wikipedia above, but it will only bring down
the error by interpolation to the order of 1m. Once you have that
accuracy, you need to evaluate the GPS data in a completely different
way (evaluating phase offset data) to go the next step of accuracy.

Rolf

On Mon, 15 Feb 2021 23:29:55 -0000 (UTC), News 2021
wrote:

If you are using 'back processing' and can afford it, for the equipment
and subscription(? dated knowledge), you can achieve in the order of 1mm/
a few mm accuracy.

Nope. Maybe 1 to 5 cm accuracy at the best of time. Post processing
is basically DGPS (Differential GPS), where corrections for a nearby
known location (benchmark) are calculated and used to provide
corrections to the data received from the satellites. Plenty of
problems here. One is that the known location might be many miles
away from the area being surveyed. Put a weather front in between the
receiver and the benchmark and the corrections go insane for a while.
Corrections make the assumption that atmospheric corrections are the
same for both identical at both locations.

For the old system run by the USCG, the nearest DGPS stations and VLF
transmitter is at Pigeon Point CA, about 70 miles away. Performance
was bad enough that MBARI (Monterey Bay Aquarium Research Institute)
had to install their own system on Mt Toro and broadcast corrections
on their UHF commercial channel. WAAS arrived just in time to allow
the USGC to shut down most of their VLF beacons and recently turn off
the original DGPS abomination.
https://www.navcen.uscg.gov/?pageName=dgpsMain

WAAS does DGPS in the same way as the older DGPS systems except that
it transmits corrections via multiple satellites frequencies at
approximately 1.5GHz. This is close enough to the GPS frequencies to
allow them to share the same antenna. WAAS is now built into most GPS
chips and quite common in today's smartphones. WAAS improved GPS
accuracy from 20 to about 2 meters.
https://en.wikipedia.org/wiki/Wide_Area_Augmentation_System

To obtain accuracies in the cm range, plenty of things need to change.
More processing time is required. Access to military channels would
be nice, but probably isn't going to happen. Antennas need to
insensitive to ground and building multipath reflections. Something
like these choke ring antennas:
https://www.google.com/search?q=choke+ring+gps+antenna&tbm=isch
The indicated location is inside the antenna. DGPS and AGPS
(Augmented GPS) corrections need to be carefully chosen. Using
corrections from a distant or badly surveyed benchmark can produce
worse accuracies instead of better. If available, using terrestrial
GPS satellites for additional accuracy are a big help:
https://en.wikipedia.org/wiki/Pseudolite
Most large airports have these. There are probably more requirements,
but that's all I can remember. Incidentally, one interesting use for
pseudolites is at amusement parks. Instead of putting the rides on
tracks, they use road wheels and guide the ride cars with DGPS with cm
or better accuracy. Same with following the path of race cars on the
track. I had an old Autofarm tractor guidance system:
https://www.gpsfarming.com/hardware_af_rtkautosteer.php
that allowed farm tractors to plow crop rows with much better accuracy
than with a human driving the tractor. It uses DGPS with a
corrections transmitted on 900 MHz. I was never able to reliably
measure the accuracy, but my guess(tm) is about 50cm.

Incidentally, if you actually had a GPS that could produce 1mm
accuracy, the calculations would need to include continental drift,
benchmark drift, earths rotational changes every time there's a big
earthquake, data jitter reduction, and some way to produce a single
number without having to deal with a constantly changing display
output.

Hmmm... I could probably estimate how many decimal places 1 mm might
require. At the equator, 1 degree = 60 nm = 111 km. If you GPS only
read whole degrees, you would get 111,000 meter accuracy. Using
decimal degrees:
Deg Accuracy or Resolution
0 xx. 111,000 meters
1 xx.x 11,100 meters
2 xx.xx 1,110 meters
3 xx.xxx 110 meters
4 xx.xxxx 11 meters
5 xx.xxxxx 1 meter = 100 cm
6 xx.xxxxxx 10 cm
7 xx.xxxxxxx 1 cm
8 xx.xxxxxxxx 0.1 cm = 1 mm
9 xx.xxxxxxxxx 0.1 mm

To get 1 mm accuracy, I need 9 digits to the right of the decimal
point on your GPS. I added the 9th digit to help with rounding off.

Looking at my results from an averaging test on my Moto G Power phone:
http://www.learnbydestroying.com/jeffl/crud/GPS-averaging.png
I see 8 digits to the right of the decimal, which gives me 7 digit
resolution, or about 10 cm accuracy. Judging by my watching the last
few digits change, I might be able to get a stable reading out to 4
digits to the right of the decimal point or 11 meters. That seems
about right GPS+WAAS.

Surveyors realised very quickly that all they had to do is adjust the
received signal by the offset of a known LOCAL point.

Yep. So, what happens when that known local point moves over the
years? Here's a "recent" 1995 map of the local GPS Control Points:
https://gis.santacruzcounty.us/DPWScans/recordmaps/087M48.pdf
and Azimuth markers:
https://gis.santacruzcounty.us/DPWScans/recordmaps/089M19.pdf
Later maps are in the form of LIDAR aerial surveys (which use man hole
covers as markers). Continental drift here is about 5 cm per year to
the NW. Since the local property markers, monuments, and benchmarks
all move roughly together, their relative positions remains the same.
However, their positions relative to a GPS derived position drifts. Of
course, we have some local earthquake faults that move much faster and
in many different directions. 1 mm GPS accuracy isn't going to help
much when the playing field changes by much greater amounts.

Since the offset varies over time, Initially, you could download tables
of the offset for every 5(?) minutes of a particular day. Now, with the
correct equipment, you tune your equipment to the live back channel(sub
carrier on various AM/FM/?) broadcast to have 'instant' corrected data.

That's fairly close to what is actually being done. The data is time
stamped with 100 nanosecond resolution. I'm not sure of the actual
number or accuracy.

Averaging over time for a day/week/month, with a PC is good.

I've been using various programs from Visual GPS for many years on
Windoze:
https://www.visualgps.net
Most are free and well worth using. I suggest Visual GPS View.

I've also manualy adjusted from the downloadable tables, but I've never
had the use to justify the cost of the live back channel gear.

I did that once, and never again. Too much work to do by hand.

These days, streaming DGPS RTK (real time kinematics) correction data
is commonly available via the internet:
https://www.gpsworld.com/finally-a-list-of-public-rtk-base-stations-in-the-u-s/
For example:
https://www.trimble.com/trs/findtrs.asp
https://www.unavco.org
For northern Calif:
http://seismo.berkeley.edu/bard/
http://seismo.berkeley.edu/bard/realtime/
Mo
https://www.google.com/search?q=public+rtk+base+stations
--
Jeff Liebermann
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558

On Wed, 17 Feb 2021 11:18:41 -0800, Jeff Liebermann scribed:

On Mon, 15 Feb 2021 23:29:55 -0000 (UTC), News 2021
wrote:

If you are using 'back processing' and can afford it, for the equipment
and subscription(? dated knowledge), you can achieve in the order of
1mm/
a few mm accuracy.

Nope. Maybe 1 to 5 cm accuracy at the best of time.

Shrug, obviously your experienced differed from the GovCo work I
supported.

Post processing is
basically DGPS (Differential GPS), where corrections for a nearby known
location (benchmark) are calculated and used to provide corrections to
the data received from the satellites. Plenty of problems here. One is
that the known location might be many miles away from the area being
surveyed.

Then it is not 'local'.