Hello, we have a Grove GPS and have been testing it out for our purposes. There are a few things that we noticed though, one of which is that it doesn't return the same results from day to day. Given that there are many factors in this (time of day / number of satellites, cloudiness, trees, etc.), a few questions come to mind:
Is there any way to calibrate the GPS in the program? I did not see any info on the Wiki about calibration, but I figured it's still worth asking about.
Related to #1, how do you determine how many satellites the GPS is connected to? (I had asked another GPS question about parsing longitude and latitude, and jboyton modified xuid's code so that it worked with the Grove output. Is there a(n easy) way to add to this logic, to parse the number of satellites?)
Would using multiple GPS sensors and averaging the values help the accuracy of the GPS? Or if they are both within a foot or two of each other, would they just return the same value more often than not?
Are there any other GPS receivers that you guys would personally recommend? We're trying to use GPS for outdoor navigation within a few feet each time. There are no roads or paths to "stick" to, unlike for a car GPS.
Do you guys have any other wisdom related to this to be aware of? (You never know until you ask!)
Update:
6) Does the orientation of the GPS sensor affect the accuracy of readings?
GPS just isn't that accurate, depending on exactly what you mean by a few feet. This page shows you the reasons for position error GPS error. It's suggesting that you can expect accuracy between five and fifteen meters.
So, your device is likely as calibrated as it can be right from the factory. I don't think multiple units in the same place will buy you anything either.
The GSV sentences will tell you how many satellites the GPS can see. Won't help you with accuracy though other than to tell you that it's likely to be worse because too few satellites are in view.
wildbill:
GPS just isn't that accurate, depending on exactly what you mean by a few feet. This page shows you the reasons for position error GPS error. It's suggesting that you can expect accuracy between five and fifteen meters.
So, your device is likely as calibrated as it can be right from the factory. I don't think multiple units in the same place will buy you anything either.
The GSV sentences will tell you how many satellites the GPS can see. Won't help you with accuracy though other than to tell you that it's likely to be worse because too few satellites are in view.
Thank you for the response. If error is present when recording GPS values along a route, and then the route is attempted again, but this time there is a different amount of error, would it make sense for the GPS readings to all be off by the same amount from each recorded point? For example, if the starting GPS value is 81.0425 and the next time the route is run the value 81.0437 is returned, the difference in error is +12. Assuming these "routes" would not take too long to complete (no more than a few minutes), would it make sense to take the previously recorded GPS values along the way and simply add 12 to them in anticipation of approaching those coordinates (of course, judging and applying the difference in error for longitude and latitude individually)?
I can't say I know anything about GPS, but a friend who does tells me that accuracy improvements depend on how much post processing one is willing to do.
One thing you can do is easily is enable WAAS and SBAS. I believe that that takes the approach you're theorizing about, at least to some degree. But I really don't know what I'm talking about.
You should do some reading. Here's a starting point: