(Apologies if there are lots of threads like this, but a quick search didn't return a definitive answer.)
I'm looking for an 'Arduino-compatible' GPS module with the highest accuracy possible. If it can be used indoors that would be super. Power usage & monetary cost aren't of particular concern.
What GPS chipsets should I be looking at? The SiRF range of chipsets seem to be fairly popular; the SiRF STAR III is easily obtainable, but I haven't seen the SiRF STAR IV as much.
What considerations are there for connecting a GPS receiver to an Arduino? Does it require any additional electronics on a shield/protoboard?
GPS is line of sight - some visibility to the sky is needed.
Funny, I've had GPS modules work well from inside my house with no direct visibility to the sky? And that was a surplus module from over 6 years ago or so and I suspect the more modern modules have even better sensitivity.
I have used the sparkfun module (LS30001) it is very accurate. Down to about 6". There is a library called TinyGPS. that will get and parse the values into an object so you can use them within your sketch. I have an example sketch that uses this library to find the distance between two points on earth using the harversine formula.
CrossRoads:
GPS is line of sight - some visibility to the sky is needed.
GPS signals are just radio waves, so line of sight isn't required. Of course they're not particularly powerful radio waves so line of sight certainly helps
ApexLogic:
I have used the sparkfun module (LS30001) it is very accurate. Down to about 6".
The MTK3339 (Adafruit link) quotes 3m accuracy without SBAS & 2.5m with, whilst the uBLOX quotes 2.5m accuracy without & 2m with. Additionally the MTK3339 has a built-in ceramic aerial which isn't convenient for me as they have to be oriented pointing upwards whereas I will most likely have the device on its side, whilst the uBLOX breakout has an omnidirectional antenna.
Also, though not useful for my particular application, the uBLOX operates up to 164,041ft whilst the MTK3339 maxes out at 60,000ft. Turns out this is important for people sending weather balloons & rockets up to the edge of space.
Whatever that means. I'm going with the assumption that most GPS receiver manufacturers probably quote performance using similar metrics, so if the uBLOX claims 2.5m/2m then it is probably safe to assume it performs equally or better to the MTK3339 3m/2.5m. Probably...
sbright33:
Can anyone explain how accuracy is defined for GPS? What does 3m mean?
GPS accuracy is a complex subject, there is no simple answer to that question but there is much information on the web on this subject. 3m would mean an accuracy of +/- 3 Meters.
Let me try to be more specific with my question. If I take 3600 GPS data points over an hours time, the average will be within 3m of the correct value as determined by a good map. Usually. I have done this experimentally. Each time I start over, the resulting average is within 3m. This is closely related to accuracy.
What % of them will be within 3m of this average? What product specification is used to express this? This seems to be a useful thing to know about a product before you buy it. Does anyone know what this number is called? Precision? Can you show me an example that answers my question?
Let me try to be more specific with my question. If I take 3600 GPS data points over an hours time, the average will be within 3m of the correct value as determined by a good map. Usually. I have done this experimentally. Each time I start over, the resulting average is within 3m. This is closely related to accuracy.
What % of them will be within 3m of this average? What product specification is used to express this? This seems to be a useful thing to know about a product before you buy it. Does anyone know what this number is called? Precision? Can you show me an example that answers my question?
I'm certainly not an expert on GPS accuracy, but from what reading I have done I know it's not just a matter of a simple calculation or device specification. The number of 'birds in view' at a give time seemed to be one factor I recall. I know there is a method called differential GPS where you compare a gps reading with a known stationary position value to gain an additional 'correction factor' to use in further calculations. My point is that GPS accuracy is in my opinion a very complex subject and the answer is not solely a matter of a datasheet specification stated on a specific manufacture's product.
I think you're right Lefty, it is complicated. I can measure this experimentally, but it depends on the view of the sky as well as the position of the satellites. Let's assume it's a wide open view, and that satellites are about the same during each whole hour. Although not a specific minute of that time. The only other variable is the GPS unit and chipset. My results vary widely when you change only this variable. How do you know which units are better than others, if the only thing that matters is precision?
sbright33:
I think you're right Lefty, it is complicated. I can measure this experimentally, but it depends on the view of the sky as well as the position of the satellites. Let's assume it's a wide open view, and that satellites are about the same during each whole hour. Although not a specific minute of that time. The only other variable is the GPS unit and chipset. My results vary widely when you change only this variable. How do you know which units are better than others, if the only thing that matters is precision?
I hear what you are saying/asking, I just don't have a useful answer for you. How would one looking just at the manufacture's GPS module datasheet, compare total/best accuracy between different manufactures offerings?
The joke in many engineering circles is that the manufactures marketing department writes up the first page specifications and the real engineers spend all the rest of the pages giving all the gotchas, caveats, and conditions one has to meet or assume to actually meet the page one specs. Specifications always have to be taken with a grain of salt until tested or research carefully and to be sure all manufactures are using the same assumptions and conditions, which they rarely are. Perhaps user reviews is another method to compare? I know SparkFun allows user/buyers reviews and comments on their products.
@cjdavies- The Adafruit unit in my link above works with an external active antenna. Can anyone beat this unit? Features? Performance? I need the best GPS for my next project. I just received this one.
