General GPS questions

[Glaurung]

Hello,

 

I am a new GPS user (just bought a Garmin etrex) and before using it for hikes and geocaching, I played with it for a couple days while walking with my dog. I have a couple simple questions, and most of them are not related to the etrex, but are general questions about the GPS system. I hope this is the right place to post my questions :

 

1) I live in town, and I've noticed it is impossible to turn the unit on and get signal in the center of town, but I need to go in a park with clear view of the sky. Well this is logical somhow : in the center, the buildings are too high and not enough satellites are visibles. But If I get the signal in the park, I can then walk in the town center, and I still have good signal. So why can my receiver still track the satellites in the center, but cannot locate them if I turn the unit on at the same location? And this leads me to wonder about GPS used in cars : How can the system be reliable in a town where often most of the sky is hidden by buildings?

 

2) I tried to take my GPS receiver in the train (big large window), but without success, is this normal?

 

3) While walking in the forrest, I sometimes lose the signal (normal, there are many leaves). I noticed that to restore good reception, I have to stop walking. Why is the unit not able to retrack the satellites signals unless I stop for like 5-10 s? This is rather annoying, because if you are not constantly watching the screen, the unit may stop tracking your walk, without you noticing it. This is of course not important on a "localisation" point of view: If you are not watching the unit at a given moment, it means that you don't need to know your precise position at taht moment. But if you want to use the "odometer" function to calculate the distance you have walked, you need not loose the signal.

 

4)Garmin related : If for some time the signal is lost, how does the software take this blackout into account to calculate the total distance traveled? Will it assume you went in a straight line during the portion without coverage and add the shortest distance to the counter, or will it think you were teleported from one point to another and not add anything?

 

Thanks in advance for the info...

[+JFDavis (Orion 6)]

I'm pretty sure the answer to all your questions revolves around a central theme-the almanac. A new receiver (or one that's been off for some period of time) must first refresh it's internal schedule of what satelites are visible and where they are before it can locate itself. The more interference you have during this acquisition, the longer the process requires. Once completed it can often track for some period of time without much new data. The process is essentially continuous, but some 'gaps' are expected and normal. However, when the 'gaps' become too numerous, then it stops knowing where it is and starts searching all over again. Buildings, large metal surfaces, roofs and trees overhead all contribute to gaps; sometimes they contribute to reflections and really wild positions are reported. Sitting in the clear is the solution. I haven't played my an Etrex in a while. I believe it only calulates distance when it actually has position and track-so yes it thinks you teleported across loss of signal/track gaps.

 

Hello,

 

I am a new GPS user

 

...

 

Thanks in advance for the info...

[Glaurung]

Thanks for the answer.

 

From what Ihave read about GPS, almanach information is a 37000 bits sequence transmitted at 50 bits/seconds. It takes therefore about 12.5 minutes to transfer. The receiver needs the almanach info from each satellite to determine its precise position at time t. That explains why you may have to wait up to 15 minutes before being ready to navigate the first time you turn on your receiver.

 

But what I don't understand is why I need to stand still for the receiver to re-locate the signal in case it loses it. The inability to re-locate the signal, added to the fact that traveled distance is computed only when the receiver gets a good signal makes the odometer function totally unusable, unless you constantly watch the receiver to see if it still receives a good signal. And when I am hiking, I prefer keeping an eye on the beautiful panorama. I know that the odometer is far from being the primary function of a GPS receiver, but it is interesting to be able to know how many km you have travelled and so on, at what speed and so on.

[+JFDavis (Orion 6)]

I'm not sure you actually need to stand still; more likely it's keep clear (of obstructions). I'm not sure what the time-limit is on breaking track-that is, what interval might cause a skip in odometer readings/samples. I suppose it could be determined empirically.

[+BigLarry]

I have an eTrex Vista and use it in many environments: car, biking, geocaching, city, woods,...

 

I will generally only loose signal on the north side of steep wooded hills and the very middle of cities with very tall buildings.

 

It should be noted the eTrex uses a patch antennae that works best with the unit face up to the sky, so you're looking directly down at it to read it. If you toss it in a pocket or pack and it's facing another way, the reception can be greatly affected. So it could be that when you move, you put it away and it's oriented differently with poor reception. On the bike I put it on the handlebars flat face up. When hiking, I have it mounted on top of a hiking pole face up. For geocaching I just hold it face up while navigating to the cache.

 

When using the eTrex in the car, I find I need to get it close to a window to get reception. Up on the dashboard is best, but I can get signal next to a side window if there's not lots of other obstructions around.

 

If I loose signal, a new "Active Log" track is made when it recovers the satellite. So you'll have lots of active log tracks on your system. When you save, it can combine all the logs into one track depending on the period you select to save. I just clear before a trip and then save all tracks into one.

 

The saved track will record the missing section as a straight line between the points and add the distance into the total distance of the track. So if you had did a lot of zigs and zags, it will be off a little. However, I've found the eTrex will not accumulate miles or altitude climb on the odometer and elevation pages without a signal. Fortunately, I don't often loose signal, and then only briefly, even in thick trees. So this is not a big issue for me.

 

These days, I also carry a Garmin 60CSx with me to get the very best tracks on rides. It never looses signal even in the worst conditions of trees and hills. It can sit on the floor of my car and still get reception. It's much better than the eTrex. Nevertheless, I find the eTrex is more rugged and I use it for navigation on my bike and it works quite well for me. If the eTrex gets a good signal, the accuracy is about the same as the 60CSx anyway. The tracks practically overlap.

[Glaurung]

Thank you.

 

This is very useful information. I indeed stored my etrex in my shirt pocket, so it was resting vertically when I wasn't directly watching it. I will try to put it in the top pocket of my backpack to see if reception improves.

 

This is interesting to note that, although the odometer will not take the lost portions into account, this can be solved by saving the track. This can lead to some errors if there were lots of zigzags during signal loss, but it will still be a better approximation than simply disregarding portions of the track without signal...

[+BigLarry]

Thank you.

 

This is very useful information. I indeed stored my etrex in my shirt pocket, so it was resting vertically when I wasn't directly watching it. I will try to put it in the top pocket of my backpack to see if reception improves.

 

This is interesting to note that, although the odometer will not take the lost portions into account, this can be solved by saving the track. This can lead to some errors if there were lots of zigzags during signal loss, but it will still be a better approximation than simply disregarding portions of the track without signal...

 

Say I do a typical 16 mile ride. and I loose signal during 2 miles (a worst case that happens maybe once a year, usually on a steep north facing hill also with lots of thick trees). Now if I zig and zag a bit, that 2 miles of trail may be 1.5 miles of straight line.

 

So the GPS odometer will read 14.0, instead of the 16.0 I see on my bike's mechanical odometer. It lost the entire 2 miles. If I climbed during that time, my elevation gain also failed to accumulate without signal reception.

 

But when I save the track into one (which loses time information, incidentally), then I see that track's distance is 15.5 miles. The GPS odometer is still 14.0. It wasn't fixed by saving the track. You just look at the track distance instead.

 

Also, my eTex on average reads about 5% lower miles than my bike odometer. The 60CSx comes a bit closer, maybe only 2-3% lower. I think this is because the GPS smooths out all the little turns the bike picks up. The 60CSx just has more agressive software to catch turns, I think. A minor effect anyway.

[+apersson850]

Since this thread is intended to give answers to a phenomena, where the answers just have to be pretty technical in nature, I'll take the liberty of jumping in to sort out some errors, which appears above.

I'll still try to keep this fairly simple, so GPS purists do not need to intervene, since I know from the beginning that I'll skip a few steps.

 

Anyway, let's begin with almanac data. This is information, transmitted by all satellites, for all satellites, about their whereabouts above our heads. It takes a considerable time to transmit all of the almanac, since it's not just limited to 50 bits/s, but should also share this information channel with everything else, that's supposed to be transmitted to the receiver.

Fortunately, the almanac is just a coarse description of the satellites trajectories. Hence, it has a lifetime of several months, so once your receiver has been initiated with this information, it doesn't have to download it again, to become useful. It's sufficient that the information is gradually updated, while you use your GPS.

You can see that you have almanac information available from the fact that the skyview displays the satellites distributed over the sky. Without almanac data, satellites are shown at the horizon due north.

If you hold the PAGE button while you turn on an eTrex Vista, you clear the almanac, so if you like to, you can experiment with that, and then watch you GPS build the data tables for you again, if you look at the skyview for a few minutes.

 

Then there's something called ephemeris data. This is precise orbital data for a satellite. It's transmitted by every satellite for itself, so you need to find a satellite before you can get this data. Without ephemeris data, a satellite can't be used in the PVT calculation. (Position Velocity Time).

This data is transmitted in three different blocks, each six seconds long. The blocks start every tenth second. This implies that to get this data as quick as possible, you must have uninterrupted reception for 30 seconds. It doesn't matter in which order you get the blocks, or if you miss one block. You get a new chance 30 seconds later.

Once ephemeris data has been downloaded, it's enough to be able to decode the satellite signal for one second, now and then, to use it. This explains why getting the initial lock is much harder than maintaining it.

On the Vista, you can see that ephemeris data has been found for a satellite, since the satellite icon, and signal bar, gets dark.

Ephemeris data typically gets outdated in an hour or two, which also explains why a GPS, which has been turned off for five minutes, can regain the position in a few seconds, but it takes at least about 40 seconds, if it has been turned off for a couple of hours.

 

Now, to be able to lock on to a satellite signal to begin with, is a neat mathematical task. I'll have to simpliy this rather a lot, but there's a lot of info on the internet about this, if you want to understand all details.

Anyway, all satellites are broadcasting on the same frequency, hence their signals mix with each other. To make matters even more complex, the signal strength is a good bit below the average noise level. So for the GPS receiver to sort out the signals from the satellites is much like being able to dechipher what twelve persons are whispering, in the crowd listening to a rock concert.

To make it at all possible, all satellites send a code, which is a mix of ones and zeroes, in a seemingly random configuration. It's actually not random at all, so it's called pseudo-random. The different satellites have their PRN identifications, Pseudo Random Numbers. The GPS receiver knows very well which seguence of bits a certain satellite is transmitting, from its PRN. By generating the same sequence internally, and then adding it to the received signals, the GPS is trying to match the internal code with the transmission. The code takes a certain time to transmit, and time equals distance, so it's essential to see "where along the code" you are located.

If you imagine two combs, where you've broken off a pins in a certain configuration, equally on both combs, but offset in length, then you can move the combs along each other, while looking through the pins. Call one the master, the other the slave. In all positions, there will be holes where pins are missing in your master, since by coincidence it will line up with holes in the slave too. But at a specific position, you'll suddenly get exactly the same pins and holes, as you made. Then you've found where they match.

When looking for a satellite, several of the transmitted bits will be decoded wrong, due to interference, so you'll not get a perfect match. But at some instance, there will be more than 50% holes where you want them, and the same for the pins. Then you've probably aligned you internal code with the satellite. This process is called correlation. To speed things up, there are more than one correlator working on the same channel, so the GPS is making several assumptions simultaneously, and then testing how good they are. SiRF Technologies use many more correlators than other manufacturers, which explains why they are so much better at locking on to the signals.

The satellites are travelling fast over the sky, but this can be predicted from the almanac data. However, your moves can't be known until this correlation has been established with enough satellites to compute your position. Hence, if you are running around when the receiver is trying to lock on to the signal, it will get more confused, since what seemed true a moment ago, is no longer true now.

 

To compute your position properly you need four satellite signals locked at the same time. If you have ephemeris data established for more satellites than that, then chances increase that the unit at any time can find a subset of these, with enough satellites to establish your position. If it can't, a Garmin will assume the last known speed and heading for up to 30 seconds, before it resigns and tells you that it has lost track of where you are. As long as you don't loose the signal for more than these 30 seconds, it will assume a straight line from the last known position, to where it figures out that you are now. But if tracking is lost all together, then it has no idea at all about where you have been, and this part will not be taken into account. If you save a track, then that track will skip over the missing parts, with a straight line.

 

Modern trains often have metallic coatings on their windows, to reduce radiated heat into the carriage. This coating very effectivly attenautes radio signals in the frequency band used by GPS.

 

Some navigators for automotive use are augmented by inertial navigation. If GPS signal is lost, then the gyroscope and wheel sensors are used to calculate turns and travelled distance. In any case, a good antenna on the car roof improves the situation, since it has an unobstructed view of as much of the sky as possible, and also has a very good ground plane in the steel plate of the roof itself.

 

I hope I've at least contributed to resolve some errors in the previous posts, and perhaps explain some still open questions, in spite of the heavily simplified description of some of the issues covered here.

Edited September 9, 2006 by apersson850

[Glaurung]

Since this thread is intended to give answers to a phenomena, where the answers just have to be pretty technical in nature, I'll take the liberty of jumping in to sort out some errors, which appears above.

I'll still try to keep this fairly simple, so GPS purists do not need to intervene, since I know from the beginning that I'll skip a few steps.

 

Anyway, let's begin with almanac data. (...)

 

Well many thanks for this in-depth explanation. I now understand better how the whole system is working. As an engineer (although this is far from my field), I like to know how things I used are working, and Garmin user manual was very poor in information.

 

Modern trains often have metallic coatings on their windows, to reduce radiated heat into the carriage. This coating very effectivly attenautes radio signals in the frequency band used by GPS. (...)

 

I was suspecting that kind of probelm. I will have to give it another try when I take a regional train without tainted windows. As for coupling GPS and an IMU for cars, I knew busses in my town had a similar system, but was unaware that it was included in wide market automotive GPS receivers.

 

I hope I've at least contributed to resolve some errors in the previous posts, and perhaps explain some still open questions, in spite of the heavily simplified description of some of the issues covered here.

 

Yes, thank you once again for the time you took to write your very clear explanation. If you have websites or book to recommend me to learn more about the subject, please tell me.

[WR8Y]

I have had problems in heavy tree cover, too. I use a Lowrance Expedition C, known for having a good recieve, but it STILL has challenges in heavy trees.

 

Click here and start reading at post 20 or so... and see how I dealt with it.

 

(Oh, and that was a very interesting exposition on how the sat's 'educate' the GPSr a few posts back.)

[+apersson850]

When you've read everything that's written here, then you know quite a lot.

 

If you remember it all.