An exciting update from the Swiftnav team, whose successful Kickstarter project is bringing high-precision (cm-level) RTK GPS to an affordable level (sub $1,000). First up: ArduPilot support!
From today's update:
The complete signal processing chain, from the GPS radio waves that arrive at Piksi's antenna, to the centimeter-accurate relative positions that are the system's output, has now been fully implemented and verified.
Ian says these are the covariance plots of the ambiguity resolution engine, but all I see are some cool circles.
There is still work to be done to port that code to the Piksi hardware. This is a decent chunk of software development, but we're now in the homestretch. Software development timelines are always difficult to estimate, so we want to make sure we avoid pushing the delivery date again by being too optimistic about the schedule. With this in mind, we're promising you a ship-by date of April 15th for the Production RTK Kits.
A sunny San Francisco afternoon, our dual antenna setup taking in the rays
We've added a new member to our team, one Niels Joubert of Stanford's computer science department. He's an amazing software developer and engineer, and we're very lucky to have him making contributions to the product. Some of you will be pleased to hear that his first project is adding support for Piksi to the Ardupilot system. This support is an utmost priority for us, and now we can finally give it the resources it deserves.
Niels Joubert / Piksi / 3DR Pixhawk
We've now finished shipping all the SUPPORT and T SHIRT rewards, so if you haven't received yours yet, it should be arriving soon.
Thank you for your continued support and patience!
Colin and Fergus
Comments
I'm thinking this will be great for net landings, like possible with the piccolo autopilot.
I've heard there's a lot of concern about EMI from UAV componentry impacting the accuracy of the RTK GPS system. Any word on this? I've been following this work for a long time, but have yet to hear any UAV application accuracy results.
It's good to see that the ambiguity resolution engine appears to be performing well in the covariance plots. What I see is a lack of variable correlation (the two variables fluctuate randomly) with normally distributed (aka Gaussian) noise. I assume this is for the covariance of lat and long positions calculated by the fast ambiguity resolution algorithm. You would expect the two to have random noise when the GPS unit is not moving for a lot of reasons. More information about this sort of thing can be found here: http://ocw.mit.edu/courses/earth-atmospheric-and-planetary-sciences... and fast ambiguity resolution algorithms here: http://www.navcomtech.com/navcom_en_US/docs/download_center/white_p...
Can I ask whos carrier phase ambiguity resolution algorithm is used?I know these patents are defended to death.....
Just incredible to implement this technology in our UAV...
Cost seems to be high for early adopter but for having such an accuracy: centimeter !!! Wahooo
Future is already there ?
Awesome effort guys!
Is this a plug in component on the APM 2.6 copter?
Would it use the telemetry channel for CMR correction signals or would it be able to receive over GSM?
You are right Marc, sorry for the confusion!
It will be great to see what it can do!
Jesus,
you are confusing differential GPS and RTK. It has something in common, but to be general, RTK is a method to massively improve precision by measuring carrier phase signals rather then the code, which is normally done. Since the phase returns every 20cm or so (sinus wave) you have to know, "which sinus you are on". For that, you use information given by the fixed ground station. (very basic description, soryy, there was a good link here recently).
All that principle makes the whole RTK system very sensitive to movement of the receiver (phase jumps) and to bad constellations of the satellites, which I would include urban environment with...
Even with a professional, >10k$, L1/L2 system I have times in fully open environment when you can´t get a RTK solution. At least these systems tell you and you can predict by looking at the constellation.
So to sum up: RTK is nice, but you have to undestand it and know the limitations!
$1000 is the new affordable.
what will be the price|>:
It is very nice! Centimeter precision is just the next step needed for APM to become a professional tool for agriculture, etc.
However. Does it improve GPS accuracy in urban canyons (mainly affected by multipath)? As far is I know about RTK, the base station measure the errors in GPS received signal by knowing a-priori it current well defined position. Those errors are broadcasted to the remote RTK GPS which uses them to correct its own observed position.
But in an urban canyon, the multipath seen by base station may be very different from that observed by the remote station due to the complicated surfaces of buildings all around...
-
1
-
2
-
3
of 3 Next