I’m currently working with EKF-3 on ArduPlane 4.5.5 and have run into a couple of issues that I’m hoping to get some help with:
- Altitude Glitches: During my flights, I’ve been noticing some altitude glitches that I suspect are due to EKF resets caused by excessive innovation. I believe this could be related to the need for better tuning of the Kalman filter parameters, but I’m not sure how to approach this effectively.
- Replay Tool: I’ve also tried using the Replay tool to fine-tune the EKF parameters offline, using .bin files from actual flights. Unfortunately, I haven’t been able to get it to work, which is preventing me from testing different parameter configurations in a SITL environment. The issue has been reported here too: Log replay generating erroneous log file with no effect of parameter change
I’m wondering if there’s a tool or method out there that would allow me to test different parameter configurations using raw flight data.
I’ve attached two graphs below: one showing the vertical position innovation and the other showing the vertical position, where the jumps are clearly visible.
Any advice or suggestions would be really appreciated!
Hi everyone,
I wanted to provide an update on my situation. I have tried modifying the parameters both individually and by loading a .param file with the settings I want to test using the --param-file option.
However, when loading the file, I encounter an error with the last parameter, which prevents the process from proceeding.
Starting sketch 'Replay'
Starting SITL input
Using Irlock at port : 9005
Starting
Home: -35.363262 149.165237 alt=584.000000m hdg=353.000000
Init GCSDummy V3.1.4-dev
Free RAM: 131072
Booting 3/1
0 0 0 validate_structures:514: Validating structures
TOGCS: ArduPilot Ready
Failed to set parameter YAW2SRV_SLIP to 1.000000
Replay counts: 0 bytes 0 entries
Even after removing the last parameter, the error still occurs with whatever the last parameter in the list is. I have checked for any formatting issues, but the file seems to be intact and correct, so I would rule out formatting problems.
The Replay tool is able to correctly extract the parameters from the .BIN file that I used to record the flight, but I am unable to modify them in order to test different conditions using real-world data.
At this point, I would like to ask if there is another method to test ArduPilot code using raw data, or if anyone has suggestions on how to resolve this issue.
Thanks in advance for any help or advice!
