Prepping to fly downtown

[Brojon]

So of course I get a lot of locals preaching gloom and doom wanting to fly downtown around some of the high rises. Biggest obvious issue is the wind that gets funneled between some of them. I have an anemometer to measure those and mark them to avoid or fly with not against.
But they keep preaching about multipath distortions and reflections mucking up the GPS and control signals...
Being an electronics and communications trained person I'm not convinced - but neither am I an unbeliever.
I was wondering how I could use my Mavic to collect any communications data to check on later. Is it the launch that starts the log? If I literally hooked everything up except the props and launch from the app - would it collect data even without control input like forward stick and all that?
At its core I want to simply walk around with the Mavic and have it log the path and telemetry data to "replay" later. Any ideas how I could do this?

 

[AlanTheBeast]

So of course I get a lot of locals preaching gloom and doom wanting to fly downtown around some of the high rises. Biggest obvious issue is the wind that gets funneled between some of them. I have an anemometer to measure those and mark them to avoid or fly with not against.
But they keep preaching about multipath distortions and reflections mucking up the GPS and control signals...
Being an electronics and communications trained person I'm not convinced - but neither am I an unbeliever.
I was wondering how I could use my Mavic to collect any communications data to check on later. Is it the launch that starts the log? If I literally hooked everything up except the props and launch from the app - would it collect data even without control input like forward stick and all that?
At its core I want to simply walk around with the Mavic and have it log the path and telemetry data to "replay" later. Any ideas how I could do this?

I think it's on motor start that the log starts. Not sure about the internal log. (.DAT files). So, sure with the props removed (and people calling 9-1-1 to get the rubber-walled truck to pick you up), you could record the data (and video).

It would be cool if you could do your walkthrough with a separate GPS receiver and see how that behaves. Plot the position record to see how bad it goes off from the route in Google Earth or some such.

You can be sure that GPS suffers from urban canyon issues. So if the home point is badly recorded (and that's likely as you're at the bottom of the canyon where it is worst), you could get into an RTH that goes to the wrong place (ie: inside a building). Or the infamous position correction fly aways.

GPS position is based on the distance from the satellite to the receiver. 4 or more GPS' sats give you 3 intersecting spheres and the 4th sat corrects clock bias (receiver). In reality, it's the intersection of 5 - 10 or more spheres. (This is called "multilateration" by the pseudo-range from the satellite).

But if some signals are blocked and others arrive via multipath, then there is less "good" info and more "bad" info as the reflected signals have a longer length (time) so that pseudo-range looks longer to the GPS. Moves the position. And some will have several multipath signals for the same sat resulting in an erratic solution depending on how the correlators are behaving that day.




Below you see the "truth" in light blue and the GPS solution in red. (The green is a corrected model - DJI don't have that AFAIK).:

 

[geigy]

Nice illustration of the problem, with ACTUAL data, or modeled data anyway. Thanks!

 

[Brojon]

In fact Austin doesn't really have the ginormous high rises of most big cities which is where I think the locals get the idea from. Tallest building we have is like 680 ft - the rest are 550 and less and there's only a handful. There's actually Wikipedia entries that will show you the tallest buildings in a given town - pretty cool.
I have a GPS track from my Garmin VivoActive HR+ that shows it tracks pretty accurate on my walks so it can't be all that bad. I was in fact planning to use it for comparison.
My understanding of GPS is it can throw away any corrupted or data that doesn't correlate with other receiver data - this is how it keeps more or less on track.
So motor start was a good guess on my part. I just wasn't sure what it would do for data collection purposes. What the hell - I'll give it a shot.

 

[AlanTheBeast]

In fact Austin doesn't really have the ginormous high rises of most big cities which is where I think the locals get the idea from. Tallest building we have is like 680 ft - the rest are 550 and less and there's only a handful. There's actually Wikipedia entries that will show you the tallest buildings in a given town - pretty cool.
I have a GPS track from my Garmin VivoActive HR+ that shows it tracks pretty accurate on my walks so it can't be all that bad. I was in fact planning to use it for comparison.
[1] My understanding of GPS is it can throw away any corrupted or data that doesn't correlate with other receiver data - this is how it keeps more or less on track.
So motor start was a good guess on my part. I just wasn't sure what it would do for data collection purposes. What the **** - I'll give it a shot.

Have fun. Referring to [1] above, not all (in fact very few) receiver employ RAIM (Receiver Autonomous Integrity Monitoring) and I'm quite sure that DJI's consumer/mobile GPS does not. Such, with 5 sats can detect a problem and with 6 sats can eliminate the single offender. But I somehow am not confident that DJI's GPS from uBlox have RAIM. (The few RAIM GPS modules from uBlox are not the models used on DJI drones). That might change in the future as one of the RAIM capable uBlox modules also has Galileo and Beidou. So the RAIM capability may work its way over to other modules that are the ones DJI are prone to pick as well as increasing the satellite system coverage. A single drone could be receiving on the order of 40 - 44 satellites at a time for positioning.

 

[geigy]

... [re carrying a drone to scout out GPS accuracy in the city] So motor start was a good guess on my part. I just wasn't sure what it would do for data collection purposes. What the **** - I'll give it a shot.

If it's a hot day, strap on the Mavic like a hat while you walk around and it will keep you cool. No weirdness there.