nachoo said:
The GPS module is usually the same in almost all the cameras? When the camera starts, Is it looking for satellites all the time, or maybe one time each 10 minutes, for example? This could explain the problem of this post. Some people says that when the cam starts in a covered place, the GPS module takes more time to know the exact place than if it starts in a opened sky, overall because it needs several satellites. How much of them? What is the most usually used gps module in dashcams? Sirfstar, FVM11, FVm8, MT3329... Can you find an explanation? You are a manufacturer, and maybe you can tell us an example of one of your cameras? Maybe the person that has the problem can find an sollution for his problem
lots of different GPS solutions available and the operation of each is something that needs to be catered for in the firmware, using a firmware from another model is risky as it's possible to damage the GPS chipset if the parameters are wrong because the chipset is different, of course when using firmware from some other model there are other risks due to hardware compatibility issues not just the GPS chipset, even the system memory used can make a difference (we struck this problem about a year ago with Hynix memory and have used Samsung exclusively since)
it's important to understand how the GPS functions as to why it behaves the way it does, with these cams the priority is always video and processing power is allocated more for video than anything else, you'll notice Chinese cams typically boot and start recording very quickly, GPS is a lower priority process and it will hunt for a signal after a recording has commenced
there are three distinct modes for the GPS, cold start, warm start and hot start, most people would be aware that it takes three satellites to get a lock, the GPS will increase accuracy as more satellites are locked and will sometimes lock 7 or 8 satellites at once, each satellite provides it's own position information and instructions to the receiver (Ephemeris data), the download of this information takes 12.5 minutes (if it all comes down error free, otherwise it starts again), this data provides accurate forecast position information valid for the following few hours but only for that particular satellite, there is also Almanac data which is much less accurate but valid for much longer periods and for all satellites, obviously you typically get a signal much faster than 12.5 minutes and some basic triangulation and positioning info can be provided relatively quickly in some circumstances depending in the state of the GPS, cold, warm or hot
A cold start is when the GPS has been off for more than 4 hours and this process must begin from scratch each time based on Almanac data of roughly where the Satellites might be
A warm start (for most solutions) is around 30 minutes, if the GPS had signal within 30 minutes prior to a restart then the Ephemris data collected is still valid and the GPS has an idea which Satellites to look for and reasonably close to where they should be at the time, the warm start is generally quite quick in comparison to a cold start, we recently changed to a GPS solution that has an internal battery so the warm start times offered are longer than 30 minutes
A hot start is when you have a signal already and you drive out of coverage, through a tunnel, under a bridge etc and when you come back into a coverage the GPS knows what Satellites you were connected to and will look straight back to those same ones and will pick the signal back up very quickly
My comments earlier about typical behaviour about being slower to catch the signal when starting up under cover and with no signal taking longer is because when there is no signal and a cold start or warm start happens if you don't get access to open sky within a relatively short time the GPS effectively gives up looking where it thinks it should be looking and starts looking elsewhere for signal which might not always be the appropriate location so it will take a bit to get back to where it needs to look
GPS has got smarter over the last few years and there's things like forward error correction to improve cold start performance and we've been looking at things like dropping invalid data for situations like the GPS starting up when there's no signal to be had to try and improve the cold start hunt time, it's quite an involved process though and something we are still working to improve, we're certainly doing better with it than we were 12 months ago but still some room for improvement I think, I guess there's probably suppliers out there that don't work to improve this and just work with what they've got, I can't speak for what others do but I feel it's something that needs more effort applied
hope this goes some way to explaining what I meant earlier about typical behaviour