One of the problems with Reach is that the reported ground course (heading) is wrong. That means that your GPS heading and your compass heading will not match. I’m not sure if the EKF will be happy with that. The good news is that I fixed that, the bad news is that your are going to have to recompile the Reach source code in order to integrate the fix into your system.
and
I’m interested in your flight findings with reach as main and only GPS, please post them.
I do not get this one, you can get both shared telemetry and injection with a single pair of 3DR radios. Are you saying that you need two pairs for this ??? Then you probably did not configured your system right.
I am not using GPS correction injection through Mission Planner, that was my point. From the https://docs.emlid.com/reach/ardupilot-integration/
(Reach supports a number of ways to accept base corrections, including the popular in UAV area serial radios. However, having a separate radio link for base corrections only is highly ineffective.), i use the “ineffective” way
I am using one pair of radios for direct communication (correction) between the base and the rover GPS. The other i use as standard telemetry radio (i.e. pixhawk to mission planner), i hope this clarify things.
What do you mean by GPS heading, is this calculated value from reach? Why would this be included in the EKF it is calculated after all. And we have the compass for that purpose which would be more accurate in showing orientation that the GPS heading.
Because that is what the EKF does, it uses redundant measurements( multiple compass, multiple GPS headings) and finds a most probalble heading. Maybee you should google EKF to find out more about it.
Again why use a single sensor when you have a lot of sensors and a EKF algorithm ???
Did ypou understand what EFK is ?
Exactly, measurements, and GPS heading is not a measurement. I know perfectly what EKF is and how it works, and you could of course include calculated GPS heading in the orientation estimation. I didn’t know that pixhawk does that. From what i know EKF in paxhawk uses IMU data and the compass to calculate the orientation and improve GPS accuracy. But again this is my assumption.
I will test my configuration in the air and see how it behaves.
Dear sir , have you solved your problems ?
here is my setting up . hope can help you .
i am also stucked here . cant get fix , no matter what i do . cost me about 1month
but i can get corrections from base .
Rover:
RTK settings:
kinematic mode
Fix-and-hold GPS AR mode
GLONASS AR mode off
Correction Input Serial UART 38400 RTCM3
Position output
Serial
UART
38400
ERB
Basemode - OFF
Base:
Correction input - OFF
Position output -OFF
Base mode
TCP
Server
localhost
9000
Mission Planner:
All subtle parameter changes exactly per tutorial
Inject GPS:
TCP client
Port 9000
here is what i have done , i have 2 reach units
one as base , the other as rover .
base connect to pc , telemetry connect to pc too .
setting as you can see above .
the other setting about mission planner &pxihawk &telemetry , please see this for your reference .
Any ETA on when a release will be made with the proposed bug fixes? These prevent the Reach from being used as secondary GPS on the Ardupilot platforms.
Have you made any progress with confirming this works reliably versus with SBAS/DGPS? Or is there other ways to fix the primary GPS to the ubox + compass module?
I am not sure what you mean by this. I assume you consider the reach module as primary.
As i said, i use the reach module connected to the gps port on pixhawk and the compass from the original gps connected to the I2C port. (You will need to connect the power pins also to the I2C cable for this).
I tested it and it works ok in normal flight and loiter. I haven’t tested auto mode.
One thing that i noticed is that i get GPS horizon error when there is no gps fix. Probably this is fine when you are outside and have good gps coverage.
Another important thing that i need to test is what happens if the base is turned off during flight.
