UAV mapping?

Would Reach be able to duplicate what the V-Map system does?

http://www.microaerialprojects.com/v-map/

The V-Map system is substantially more expensive at about $8500 a pair so the Reach is more than interesting! Actually the V-Map isn’t even RTK as it is a post processed. We currently have one RTK system, but it’s for our lidar. We would ad the Reach to all our photogrammetry rigs at this price!

With work, I don’t see why it couldn’t.

When I suggested Reach to one of my GIS clients this was his rather terse response:
Mike
Its single frequency (ublox 8) and therefore is crap at RTK. You need dual frequency modulation for RTK and post processed to get consistent cm accuracy.

I tried setting up a single frequency RTK with a couple of years ago and it was crap. And I used Hemisphere gear to do it - $5K each. For three days with manufacturer – just unreliable.

http://www.canalgeomatics.com/product_details.php?product_id=165

I bought the Orange Altus units I use instead which have not missed a beat.

all survey grade GNSS use Dual frequency (L1 & L2 as a minimum along with L2C and L5)

also multi frequency /phase is used to remove multipath and ionosphere interference to attain higher precision

Cheers

I am no GPS expert so is what he saying true?
Is this single frequency solution limited in accuracy as much as he says?
Feedback from anyone would be appreciated.

If this is indeed true, what will deviation be from day to day, say, good weather conditions? Will I be able to set this up one day and then the next day be able to place the Rover in the exact same position and have yesterday’s coordinates very closely align with today’s? Or at least within about 6inches? It looks to me like dual frequency is the way to go, but how well will this perform matched up against it?

Thanks,
Weston

@mBoland Not true.
On short baselines there is not much of a difference in accuracy between L1 and L1/L2 (especially in post-processing). L1/L2 receivers are better in other things: they allow for longer baselines (hundreds of km) and have smaller fix time.

@kk6mrp The resulting coordinates are always relative to the base station position, if the base is in the same place and conditions are good you will get your coordinates very close. Please remember that Reach is an affordable solution for makers and hobbyists and it is not competing with $5000 receivers from big companies, their goals are different.

2 Likes

Ok, that makes sense… And, yes, the goals have their differences depending on the use.

What’s the accuracy of Reach’s RTK solution when the rover is flying with a speed, like a mapping drone(about 10-15 m/s)?
Anyone tested it maybe??

2 Likes

You’re certainly onto something here. Speed induced RTK errors I believe are the biggest challenge facing mapping applications.
Assuming 10Hz RTK, 10m/s = ~0.2m travel distance between corrections plus the system latency ie: data transmit time from base to rover.

This alone is the reason why if you look at high end commercial drones (Trimble UX-5), they don’t use RTK for mapping but rather post-process using PPK.

RTK is still used to enhance navigational accuracy though.

All survey grade gear has been L1/L2 for at least the past decade. Most now are using L2C and L5 signal as well. L5 is stronger and gives better coverage under challenging conditions.

It’s not uncommon for a survey grade receiver to fix in less than 5 seconds and have over 18 sats visible at any one time (at least in central Canada). I’ve seen 24 sats visible at times with GPS, GLONASS and GALILEO enabled on a L1/L2/L2C/L5 receiver, achieving PDOP under 0.6! That’s what a complete $50k base/rover will get you 3 years ago.

Hopefully Reach 2.0 closes the gap soon. :smile:

mBoland response above is spot on the money.

Check out these spec sheets to see what big money will get you.


2 Likes

How do you post process RTK Data for a drone while in flight?

1 Like

You’d essentially be collecting raw GPS data while in flight and post-processing it afterwards.

For example, have a base station setup collecting raw data on the ground while flying the drone and collecting raw data while in the air.
The two data sets are then brought together and post-processed. In this way, there is no induced latency of the correction signal between base/rover.

RTK surveying has a correction signal latency on the order of 1s. This is no big deal for a surveyor in the field who just holds the GPS receiver steady over a point for a few seconds to get good positional RMS location but for a moving object, RTK is less than ideal for high precision work.

1 Like

Chris, this latency does not directly translate into latency in coordinates. The main issues with UAV mapping and speed is that even at 10Hz RTK and plane flying at 20m/s you will get an update every 2 meters, which is not enough. To solve this issue Reach has timestamp pin, which can be connected to flash sync on the camera to precisely mark the time when the picture was taken. In post processing that will allow us to determine camera positions down to centimeters.

Adding time marks processing is a work in progress, but as you have probably noticed we are rolling out new features almost weekly, so this should be available soon as well.

1 Like

This is essentially the definition of PPK. Post-Processed Kinematic.
http://www.wirelessdictionary.com/Wireless-Dictionary-Post-Processed-Kinematic-PPK-Definition.html

Igor, are there any indications that you can give us about the chances of seeing an L1/L2 or other upgraded version of Reach within the next year or so? I believe that with the extensive availability of VRS networks all over the world, a budget friendly L1/L2 receiver that can take advantage of longer VRS baselines would be a huge product success.

2 Likes

Can’t say anything about that.

There is lots of misinformation about the performance of L1 systems. We have achieved great results with UAV mapping using Reach-like system, that were indistinguishable from L1/L2. On baselines under 10km you do not really need L2.

With VRS you are working with almost zero baseline. This is why they are called virtual reference stations and are especially beneficial for L1.

I know that guys selling L2 systems for $10k are doing their job very well and are very convincing telling that L2 is the only way to go. Surely, they will never tell you that at 200$ Reach can do the job as well.

2 Likes

I believe there is a need for baselines longer than 10km, nobody is arguing against the accuracy of the Reach modules, but there is a definite need for longer baselines than 10km.

1 Like

Sure, no doubt that adding more frequencies will be great! We closely monitor the GNSS world and as soon as we are able to deliver a multifrequency solution at a reasonable price we definitely will.

2 Likes

How much would it cost for a multifrequency GNSS solution at current rates?

1 Like

V-Map system starts at about $8500 so I guess that’s the bottom of the L1/L2 market.

Weston Miller, can you explain why? If you’re using a Reach RS base and a Reach UAV rover, the rover is never going to be more than a km or so away from you (assuming you fly legally within VLOS). The only issue is whether the base can get GNSS corrections. And if it can’t I assume the data can be post-processed anyway? Igor?