Question regarding GPS?

[Sam_40]

Hello,
I am thinking to build a real time, very accurate GPS unit for land surveying/mapping. Using a GPS unit (did not decide yet on what to use), PIC (did not decide yet on what to use), 4X16 LCD to display the info and maybe a max232 for the GPS if required (every thing that I found uses RS232).
I searched the forum and google. Nothing helpful as far as:
1- What is a good and accurate GPS unit?
2- What will be the best choice as far as PIC from the 18F family?
3- What kind of accuracy should I be expecting out of this?
4- Any good CCS code example to start with?
I am sorry if this is off topic.
Any input will be appreciated.
Thanks,

[temtronic]

0) overall you won't be able to compete with a commercial product, so I assume for home use...

1) get one that rcvs a LOT of sat hits, the more the better, more accurate results

2) I like the 18LF46K22 ,since it has a lot of memory and peripherals( including 2 HARDWARE UARTs). You'll need a 3V PIC as all(most) GPS modules are 3volt device, best to check the SPECS!

3) accuracy is based on sat hits, so within say 30M ? It depends on GPS unit.

4) I don't know, not having 'played ' with one.

I'd use a 4x20 display, same price as 4x16, try to get a 3 V version. Having ONE power supply makes life a LOT easier. 4x20 allows for MORE data to be presented, more 'options'. If you select a PIC with MORE features than you 'think' you need, you will save a LOT of R&D time($$$) and have the project finished faster. Consider battery life, bigger batteries last longer, do NOT try to make this thing 'micro' in size!!

hth
Jay

[Sam_40]

Jay,
Thanks for the reply:
0- Yes, it will be for personal use for now.
1- Any one in particular that is accurate and had been tried on similar project?
2- Thanks, I have couple of the 18LF46K22 chips and some others, with at least 1 uart.
3- 30 meters? Holey moley I get better accuracy with an old Astrolabe
4- Any input on this will help me a lot.
5- I only wanted to display the time and the GPS coordinates. I thought 4X16 would be enough; however I see you point and I will consider 4X20.

Thanks,

[asmboy]

there are a number of potentially useful modules
check:

http://www.adafruit.com/datasheets/PMTK_A11.pdf

drivers are another matter and you have serious technical
issues to consider going well beyond the PIC program.

You say "land surveying".....

check this link
http://www.dot.ca.gov/hq/row/landsurveys/SurveysManual/06_Surveys.pdf

In the years when the system had "selective availability" ON
accuracy was quite poor for civilian users -
and even today with single point measurement it is not so easy to
get surface coordinate repeatability to better than a CEP of ~25 ft or so
owing to multipathing, delta cosine error with altitude and doppler jitter when moving. While fine for walking or driving - it is laughable for surveying purposes.

assuming no hot air balloon hovering - and a solid static platform - two error sources are eliminated. you should GOOGLE this:

kinematic gps surveying technique


You really need an external, well placed antenna system to exploit even simple GPS. Not to mention good judgment in case design.

This is a major bit of work if you want REAL accuracy AND have no RF
or surveying experience.

The communication protocol is the least of your problems.

[Sam_40]

asmboy,
Thank you sir for the detail information.

[temtronic]

gotta wonder how good a 'tablet' ap might be ? I see 7" units for about $50 these days probably les than a good GPS module.

Jay

[ELCouz]

[ckielstra]

A MAX232 chip is not required as most modules use a 'logic level' voltage for the serial connection, often 3V so best to build your system on the same voltage.

Why not use a small graphic LCD? Some are even cheaper than a 20x2 LCD.

Ready built devices can be bought, so could you provide more info on why you want to do this project? Is it because you want better specs, save money or is it just fun to build?
Why can't you use a phone or tablet with inbuilt GPS?
You mention land surveying/mapping. This rings alarm bells for the desired accuracy. Standard GPS accuracy is about 15m. For high accuracy up to 10cm a DGPS system is used where GPS is used in combination with a local placed beacon at exact known reference location.

Also remember the difference between accuracy and repeatability. Trying to improve accuracy by taking more GPS location samples is only working up to a certain level as your receiver will often return the same location data with same error.

[ELCouz]

[Ttelmah]

There are several other bits here.

WAAS (this is a "poor man's" differential GPS). Uses a signal from a satellite to give a 'rough idea' of the local error, and improve the accuracy.

aGPS. Assisted GPS. This covers two technologies. The first is improving the speed to obtain the ephemeris A 'cold' GPS (one that has been off for a long time, or has moved a long distance), has an incorrect or out of date ephemeris. Normally the ephemeris is broadcast over 20 minutes by the satellites, which is why in the 'old days' a GPS could take this long to get it's first fix (or longer if the view of the sky was poor). Most modern GPS's can receive the ephemeris in bits from several satellites, so can get this in only perhaps 5 minutes. Once they have it, it is 'good enough' to give a rough location for perhaps 3months, so the units can then give an almost instantaneous first fix (few seconds). aGPS allows the ephemeris to be obtained in only a few seconds from the internet, so the initial fix is quicker, and more accurate. There is also though a variant of this, that includes the local error, and this covers a smaller region than the WAAS signal, improving accuracy very markedly. Suitably enabled devices can give positions to only a few feet accuracy with this. Some phones are using this variant.

Then you get to the 'best' the local DGPS system. Here you have two GPS's that talk to one another only a few hundred yards apart. You position one on a known 'basepoint'. This then reads the GPS position, and broadcasts it's error relative to the know position over some link (wi-fi etc.), to the other unit. The second unit then takes it's GPS position and adjusts this by the error from the first unit. Done carefully, this can give accuracies of only a few hundred cm.

However with all these systems, there is the problem of signal reflection. If you use a GPS based locator and you perhaps approach a tall building, with trees on the other side of you, you can get the situation where the reflected signal from the building, is stronger than the 'straight path' signal going through the trees. Suddenly the distance being measured to all the satellites in the 'tree' direction, is longer than it should be, and your GPS position moves significantly. This is why it is common to be able to get really good positions with surprisingly small errors standing in an open field, but the recorded position from something like a car, will have the erratic spike location many feet from where it should be.... Accurate systems will use secondary data to try to reject such errors. Classic is using an inertial sensor as well, and a position that moves noticeably further than the inertial daa says it should, can then be rejected.

Using a local DGPS system, in open areas, I've repeatedly seen better than 2m accuracy. You can also get more accurate still by recording the phase data from the two units, and post processing this information, can get down to a few cm. However move into a built up area and this degrades massively...

If you are in an area that supports DGPS broadcasting (America for example), then look at brandname GPS modules that support this:

<http://www.roundsolutions.com/shop/en/Modules/GNSS-GPS/Trimble/High-Precision-GPS-cm>

Though they quote 'cm' accuracy, take this with a pinch of salt, anywhere but the middle of a flat field, 'real' accuracies will be a lot worse. However these are an order of magnitude better than the cheap Chinese GPS modules.

[Sam_40]

Guys,
Thanks a lot for all the very informative information.
I wanted to build this project for 2 reasons:
1- I like to build this kind of stuff.
2- We have a big piece of land; it always required fence maintenance …etc.
I did use a smart mobile device in the past to locate few points and referenced them or compared them to the survey that was done by a surveyor. Most of my readings from the mobile device were off by few inches to few feet. My thoughts was to build one that is more accurate than the phone or car GPS. However it sound like this will be a waste of time to try to build something that is useless
I would like to thank everyone who share their knowledge and took the time to help me understand the concept of the GPS. It also helps to see how you guys approach these kinds of tasks so I may apply them on my other projects!
Thanks again,
Sam

[Ttelmah]

Key thing though his that you are not building the GPS, so get a module that helps as much as it can. As I said look at the Trimble units. These are not complete GPS's, but just the little modules, but these ones have the ability to use area DGPS signals like those broadcast in the US etc..

[Sam_40]

Ttelmah,
I searched mouser and digikey for the aforementioned product by Trimble with no luck.
It seems the BD920 is the device for me!
The company that you posted the link to is not a US based company!
Where would you buy their product in the US?
Thanks,

[temtronic]

Assuming US really meas U.S.A.
then
perhaps this is too obvious ?
You go to their website,
pull down 'locate dealer',
fill in the blanks
and
wait 3 seconds.....

It found 3 dealers for me for that product.....

hth
Jay

[Sam_40]

Jay,
Yes, Your assumption was right! US means USA,
I went to the Trimble website and I found a dealer. I will contact them on Monday.
Thanks,