Rover Drifting in Auto Mode with PX4FMU Pixhawk

Recently purchased original Pixhawk from 3DR . I have connected it to a Kyosho Axxe RC Car with 6 Ch 2.4 Mhz radio. In MANUAL Mode it runs really well . Compass, Gyro,is calibrated. The EKF is disabled. After collecting a few way points in MANUAL mode on a long open road by toggling CH7 on radio and then changing to AUTO mode, the car traces the way points perfectly the first time. But when I try to run the car again on the same way points , after maybe a couple of times of successful run , the car tends to “drift” in a unusual path .I think it not trying to go to home position because I am always starting in auto mode with HOME within a few meters right ahead of me. The way points stored show up perfectly on the MP Flight Planner page. It gives the impression that the coordinates of the drive way has changed. Is it possible that the GPS location of a particular spot can change within minutes ? What could be the issue ? We have tried tuning the steering but it is not solving the problem. Your help will be most appreciated as we need to complete this project .

@Pam,
What do you have NAVL1_PERIOD and NAVL1_DAMPING parameters set to?
Please describe the “unusual” path in more detail.
Please provide a tlog and a dataflash log of the mission with an “unusual” path.
Why do you have EKF disabled as it is far superior to DCM?
Regards,
TCIII AVD

Hi Thomas
Thank you for writing in. The NAVL1_PERIOD is at 8.0 and NAVL1_DAMPING is at 0.750. By unusual path I mean the vehicle tends to do the same mission but shifts a few meters either north or south , as if the entire “frame” of way points has been shifted to north or south. I am not familiar how to get the tlog and data flash log to post it here. Pl. advise. With the EKF enabled the vehicle takes a wider path.
Best
PAM

@Pam,
This section of the ArduRover Wiki will instruct you on how to download and analyze the tlogs and dataflash logs.
Regards,
TCIII AVD

i think it might be just bad GPS reception from satelites disapearing under the horizon ?

@Tom what differences did we saw from your T3 entry runs ? Was it like around a meter ?

@Linus/Pam,
On the 100 meter legs I had waypoints every 3 - 4 meters and I found that the Rover travelled in a convex curve from the beginning waypoint to the turnaround waypoint at the end of each leg which was very strange. Due to bad weather at the time of my T3 runs, I was unable to experiment with the number of waypoints between each 100 meter leg to see what was causing the Rover to run in a curved path.
I had at least 12 - 14 satellites visible at any given time with an HDOP approaching 1.
Regards,
Tom C AVD

1 Like

Thanks Tom. GPS was the LEA-6 or a newer Ublox GPS ?

@Pam,
The Rover used for T3 was running dual 3DR LEA-6 GPS receivers spaced about a foot apart.
Regards,
TCIII AVD

1 Like

Thanks Thomas . Will go through it and hopefully be able to upload the tlog.
Best PAM

@Pam,
Uploading a dataflash log would be helpful also.
Regards,
Tom C AVD