-
Notifications
You must be signed in to change notification settings - Fork 17k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
autotest: add test for EKF losing then refinding a GPS #17104
Conversation
Running this test is simply:
(subsequent runs can omit the build step if the code hasn't changed) |
I think we should close this PR because I can't reproduce the issue. I think this PR is trying to switch to the semi-autonomous mode too quickly. It needs to give the EKF more time. |
a162150
to
f7938f9
Compare
I disagree. The EKF is misbehaving and (a) this could happen on a real vehicle and (b) this might be a symptom of a deeper problem with EKF state. .... additionally, if the EKF needs more time then it shouldn't be returning "I have a position" when it does not. |
f7938f9
to
a279a34
Compare
a279a34
to
6e80036
Compare
The associated issue has been marked as "fixed", and this test appear to be passing, so I think we're OK to merge this one. |
6e80036
to
c7dd7ca
Compare
I couldn't merge this cleanly into 4.1 |
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in ArduPilot#17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in #17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in ArduPilot#17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in ArduPilot#17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in #17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in #17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in #17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in ArduPilot#17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in ArduPilot#17104 where the test was added
this test was relying on the EKF bug fixed in this PR. The copter was actually fully landing and touching the ground in the LAND phase whereas the original intention of this test was that it not touch the ground in LAND. See the graph in #17104 where the test was added
This test cruelly rips away the GPS from the EKF, waits for it to enter land mode (and until 30 seconds have passed). It then gives the EKF the GPS back and calls RTL instantly.
The EKF seems to start to give position back to the vehicle (or else we wouldn't be able to enter RTL), but a few seconds later the EKF decides it is unhappy with the world again (variances) and forces and EKF failsafe again.
There should probably be some sort of position-reset happening within the EKF here.