Jump to content
Sign in to follow this  
Followers 1
dave497

GPS Drift When Stationary -- Looking for Explanation

Recommended Posts

I've searched the forum and the internet trying to find a technical explanation for a phenemenon I noticed just recently when I looked at track logs from BeelineGPS: drifting GPS position when stationary, which clears up when I start moving.

 

It seems that everyone who has looked at raw GPS information (as opposed to software filtered information displayed by many software packages) has noticed this. It also seems like a phenomenon inherent to GPS, not due to poorly engineered receivers.

 

The GPS setup I've been using is Globalsat BT359 with a Pocket PC using BeelineGPS and OCN6 mainly. When I upload a track file from BeelineGPS to Google Earth, the accuracy while moving in an automobile is good enough that I can zoom in and see two parallel tracks on either side of the road if I do a U-turn and reverse my direction. But when I come to a stop my position, based on the track data, starts to drift, sometimes erratically, until I begin moving again. It doesn't seem to be related to signal quality. I've seen position drift when stationary with 11 satellites tracked and WAAS signal acquired.

 

I have read various explanations, including GPS signal bounce from buildings, signal attenuation from trees, satellite based corrections (WAAS) vs. ground based (DGPS), software filtering algorithms, and others. The explanations I've read are inconsistent with each other and none of them really explain what causes the drift.

 

What are the reasons you've heard for GPS position drift when the receiver is stationary as opposed to moving? An explanation won't make my GPS drift go away, but I'd feel better knowing anyway :D

Edited by dave497

Share this post


Link to post

Don't know anything about your specific setup. However, this behavior is characteristic of high-sensitivity receivers, particularly in environments where mutli-path reception is likely (which would be just about anyplace other than the middle of a football field). In a multi-path environment, the receiver "sees" more than one signal from a single satellite. This occurs because the signal is reflected from obstructions such as buildings, walls, vehicles, hills, tree cover, and the like. The signals arrive at the receiver at slightly different times. Depending on which signal the receiver chooses to believe for any given sample, distance from the satellite also differs. Multiply this behavior by several satellites, then factor it into the overall position calculation. The net result is that the receiver perceives abrupt shifts in position. Sometimes these can be significant -- say 40 or 50 feet.

 

Higher level software in the receiver tries to smooth this out. Generally speaking, it does a better job of this when you're traveling at a decent speed. I assume this is because your true change in position per unit time is much larger than the random shifts caused by multi-path reception.

Share this post


Link to post

Also, another possible factor is that GPSr with WAAS on are supposedly only accurate within 3 meters 95% of the time.

Share this post


Link to post

My forerunner 201 shows tracks like that when I am standing still. The drift isn't too bad but it still does drift. The forerunner does not have an electronic compass so I chalk up the drift to that when I am not moving.

Share this post


Link to post

Presence or absence of a compass in the unit has no effect on this issue. It's related purely to satellite data.

Share this post


Link to post

Don't know anything about your specific setup. However, this behavior is characteristic of high-sensitivity receivers, particularly in environments where mutli-path reception is likely...[snipped]

 

Interesting. The GPS I use has a SiRF Star III chipset which, I've read, is very sensitive. In fact, one reviewer (a GPS-phile, not a professional) suggested that it may actually be too sensitive for people who use it strictly in urban areas because it may pick up low amplitude bounced signals so well that the bounced signals pollute the readings. The test might be to record the stationary drift in the city, by buildings and compare it to drift out in an open field. If your explanation is right, drift should be much less away from anything that might reflect the signals. I'm going to give that a try, and record the GPS track data under both conditions.

Share this post


Link to post

When you are moving, one point a second will be one point in space and then another somewhere else a second away... Connecting the dots will make a nice line.

 

When you are standing still, your GPS will record hundreds of points, a second apart, at where it thinks it is. The GPS system has a 95% chance of your coord (this second) being within a 10 meter radius. So connecting the points will be a 20 meter jumble. It's not really as bad as that, especially with WAAS, but it will never be exactly 1 single point until you move away.

 

Plus, your GPS may still think you are moving for several seconds after you have actually stopped, so it may be guessing where you are instead of calculating where you are. Magellans are notorious for this. see my web page http://ray.jerome.jobs.googlepages.com/gpshints for details, but it may not apply to your GPS manufacturers design.

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  
Followers 1

×