Guys,

If you need precise GNSS for your robots then you need RTK. If you don't know much about RTK, you may lose a lot of money (and hair) on weak RTK GNSS systems. I've written a post to show the beginner what they need for a super solid precision RTK GNSS. Included are specific companies and prices -- I've not seen this guide anywhere else.

http://deepsouthrobotics.com/2017/04/20/spilling-the-beans-on-rtk/

Here's hoping it saves you time & money!

Sincerely,

Roby

Views: 1654

Reply to This

Replies to This Discussion

It's a great guide -- better than anything else I've seen on the subject, and it covers the ones we use such as SwiftNav and Emlid. Super impressed. 

Thanks Chris! Hey man, big personal thanks to you for your entire body of work over the past decade -- I've benefitted enormously (silently 'till now) from Pixhawk and this entire community.

BTW, I have a Piksi, too, but never got it working well. Should I give up, or have the software updates in the past year or so made it semi-usable?

Roby, Very interesting article.  Since 2000 I've used Novatel and uBlox receivers on various UAVs, particularly VTOL.  The early uBlox receivers were so-so with firmware changes from uBlox making us lose hair when one vehicle few ok and another would act "drunk".  Performance depended on location and time of day.  NEO-M8N is better than previous receivers, but still not very impressive. The M8P is marketed as an RTK receiver but it should be mentioned that my experience is that it performs really well even in single point mode.  My point being that it justifies its price even when it does not achieve RTK Fixed mode. One concern has been the degradation performance, i.e. how much will the estimated position jump when the system loses RTK lock.  So far, we have been pleased with the performance of the M8P.  More testing is needed, though.. 

Chris, out of curiosity I tried the Piksi again a month or so back -- performance still bad.


Chris Anderson said:

BTW, I have a Piksi, too, but never got it working well. Should I give up, or have the software updates in the past year or so made it semi-usable?

Thanks Henrik!

I'd love to hear more about the M8P's performance in tough environments. Also, can you speak to it's re-acquisition time after losing fix?


Henrik Christophersen said:

Roby, Very interesting article.  Since 2000 I've used Novatel and uBlox receivers on various UAVs, particularly VTOL.  The early uBlox receivers were so-so with firmware changes from uBlox making us lose hair when one vehicle few ok and another would act "drunk".  Performance depended on location and time of day.  NEO-M8N is better than previous receivers, but still not very impressive. The M8P is marketed as an RTK receiver but it should be mentioned that my experience is that it performs really well even in single point mode.  My point being that it justifies its price even when it does not achieve RTK Fixed mode. One concern has been the degradation performance, i.e. how much will the estimated position jump when the system loses RTK lock.  So far, we have been pleased with the performance of the M8P.  More testing is needed, though.. 

Roby, Those are good questions and I will get you answers!  However, because GPS conditions change continuously, we'll have to come up with some measurement metrics - something meaningful to compare performance to.  For example, we have to define what a "tough environment" should be...

I just captured this image from my desk inside my house.  u-center and the LED constellation display are in general agreement (North is to the right on the desk.).  Two sats above 40 dB.  That is quite impressive.  However, it is difficult so say if this performance is a result of the optimized hardware design of the board - or the M8P. 

I'm headed to the airport in a few hours but will start the effort to perform a series of controlled experiments to quantify the performance, in a few days.      

Henrik, Your board with the positional LEDs is super slick. I'm digging that! Did you design it? Is it open source / or are you going to market it?

Hey, you're absolutely right that "tough environment" isn't at all scientific. My backyard testing ground is 30mx60m with a dozen 20m trees all around -- the ComNav k501g rarely loses RTK fix. If the M8P can claim to do that, I'm ready to start testing M8Ps!

Henrik Christophersen said:

Roby, Those are good questions and I will get you answers!  However, because GPS conditions change continuously, we'll have to come up with some measurement metrics - something meaningful to compare performance to.  For example, we have to define what a "tough environment" should be...

I just captured this image from my desk inside my house.  u-center and the LED constellation display are in general agreement (North is to the right on the desk.).  Two sats above 40 dB.  That is quite impressive.  However, it is difficult so say if this performance is a result of the optimized hardware design of the board - or the M8P. 

I'm headed to the airport in a few hours but will start the effort to perform a series of controlled experiments to quantify the performance, in a few days.      

Roby, Yes x 3.  Preliminary data sheet is attached.  However, before selling it and joining your list of undesirables :), we really need to do more testing.  If you are up for it, I can get you a system.    
 
Roby said:

Henrik, Your board with the positional LEDs is super slick. I'm digging that! Did you design it? Is it open source / or are you going to market it?

Hey, you're absolutely right that "tough environment" isn't at all scientific. My backyard testing ground is 30mx60m with a dozen 20m trees all around -- the ComNav k501g rarely loses RTK fix. If the M8P can claim to do that, I'm ready to start testing M8Ps!

Henrik Christophersen said:

Roby, Those are good questions and I will get you answers!  However, because GPS conditions change continuously, we'll have to come up with some measurement metrics - something meaningful to compare performance to.  For example, we have to define what a "tough environment" should be...

I just captured this image from my desk inside my house.  u-center and the LED constellation display are in general agreement (North is to the right on the desk.).  Two sats above 40 dB.  That is quite impressive.  However, it is difficult so say if this performance is a result of the optimized hardware design of the board - or the M8P. 

I'm headed to the airport in a few hours but will start the effort to perform a series of controlled experiments to quantify the performance, in a few days.      

Attachments:

Henrik -- I'm inclined to take you up on your offer. I've got access to an Emlid Reach. Looks like we need to set up a k501g vs M8P vs REACH shown down!

Henrik Christophersen said:

Roby, Yes x 3.  Preliminary data sheet is attached.  However, before selling it and joining your list of undesirables :), we really need to do more testing.  If you are up for it, I can get you a system.    
 
Roby said:

Henrik, Your board with the positional LEDs is super slick. I'm digging that! Did you design it? Is it open source / or are you going to market it?

Hey, you're absolutely right that "tough environment" isn't at all scientific. My backyard testing ground is 30mx60m with a dozen 20m trees all around -- the ComNav k501g rarely loses RTK fix. If the M8P can claim to do that, I'm ready to start testing M8Ps!

Henrik Christophersen said:

Roby, Those are good questions and I will get you answers!  However, because GPS conditions change continuously, we'll have to come up with some measurement metrics - something meaningful to compare performance to.  For example, we have to define what a "tough environment" should be...

I just captured this image from my desk inside my house.  u-center and the LED constellation display are in general agreement (North is to the right on the desk.).  Two sats above 40 dB.  That is quite impressive.  However, it is difficult so say if this performance is a result of the optimized hardware design of the board - or the M8P. 

I'm headed to the airport in a few hours but will start the effort to perform a series of controlled experiments to quantify the performance, in a few days.      

How much is this unit - k501g? Less than USD 300 per unit?

600-900 USD in low quantity, depending on RTK output rate (price quoted last year -- should less now)


Noli Sicad said:

How much is this unit - k501g? Less than USD 300 per unit?

Reply to Discussion

RSS

© 2017   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service