I’m working on a new project in which I wish to track the lat/long of an object that will regularly move to areas that have an obstructed skyview (e.g. buildings, walls, roof extensions, etc.). I tried using the Adafruit Ultimate GPS (https://www.adafruit.com/product/746) but unfortunately found that obtaining a GPS fix in many cases was unsuccessful or took an extremely long time. As the Particle Asset Tracker uses the exact same module, I expect it to yield the same result.
I’m hoping someone from experience can recommend the best-performing GPS for this use case. I’m aiming to to obtain a GPS fix in max. a few minutes, with an accuracy of 10 to 20m.
You can use a GNSS, which uses multiple nav systems (GPS, GLONASS, BeiDou, Galileo, etc).
The Particle Asset Tracker is based on the u-blox M8, which is GNSS.
I’m not so sure about the Adafruit link.
You can increase the max acceptable PDOP, HDOP, etc, but there’s really no real benefit.
There are several technologies that “help” with multi-path errors, but it all boils down to geometry.
Even a $30,000 Survey Grade GNSS can’t calculate accurate positions with poor geometry.
Waking the Asset Tracker at short intervals during movement will help you keep track of the last fix location.
If it lands in an area with obstructions, the code can continue to wake to attempt a location fix with increasing time intervals. Solar Recharge obviously helps.
I truly hope someone here responds with a product that I don’t know about that performs well under obstructions. But in my opinion and experience, those obstructions are much more important than the particular receiver sitting under them.
While more sensitive receivers will work inside a regular wood/brick/gypsum house, any metal blocks and reflects GPS signals, meaning not only low signal, but multipath interference. That’s a tough problem to attack directly, so I would consider some lateral solutions such as failing over to the cell modem position estimate.
Yeah, for my other project I already use the cellular info to approximate the device location - and it’s my fallback plan if I can’t get a GPS module to work reliably for this project. In theory it should be sufficient for my goal, but there’s definitely an advantage to be gained with this project if I can approximate the location with a much higher accuracy level.
I’ll probably just have to build one test unit and test it in practice - I’ll definitely keep the module powered for a warm start at least.
I found this GPS receiver to be better than Adafruits Ultimate GPS receiver and it’s smaller. It can receive all the GPS satellites (GPS, GLONASS, BeiDou, Galileo, etc).
I’m sure you can get the cost down on these if you just buy the GPS module to put on your own custom boards.
Since we’re talking GPS accuracy
Here is a new uBlox GPS receiver with 25mm accuracy by using Kinematics (RTK). To expensive for @Vitesze project but its new so wanted to share it.
I got curious and it seems that this RTK implementation needs a fixed reference station for regular high-accuracy solutions and can use a roving reference station for getting an accurate 3D vector between stations.
You still need good line of sight to the satellite constellation for both to make this work though.