Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #8748 (rebased) #15854

roackb2 · 2020-11-20T08:21:05Z

This is based on the work of @chobitsfan with branch and issue #8748.
I've done the rebase on master branch and need someone to check the logic is correct.
@amilcarlucas may I kindly request for your review?

amilcarlucas

remove the temp.cpp file

roackb2 · 2020-11-20T08:32:23Z

My bad, removed it.

roackb2 · 2020-11-20T09:20:17Z

I'm a little bit confused about why the test failed, I don't understand what does the DynamicNotch mean, per the test result as following:

>>>> FAILED STEP: test.CopterTests2b at Fri Nov 20 08:45:46 2020
9938
Failure Summary:
9939
  test.CopterTests2b:
9940
    DynamicNotches (Fly Dynamic Notches) (Double-notch peak was higher than single-notch peak -25.879101dB > -27.481334dB) (see /tmp/buildlogs/ArduCopter-DynamicNotches.txt)
9941
    GyroFFTHarmonic (Fly Gyro FFT Harmonic Matching) ([Errno 2] No such file or directory: 'logs/00000003.BIN') (see /tmp/buildlogs/ArduCopter-GyroFFTHarmonic.txt)
9942
    Replay (Test Replay) (Command '['build/sitl/tools/Replay', 'logs/00000009.BIN']' returned non-zero exit status 1.) (see /tmp/buildlogs/ArduCopter-Replay.txt)
9943
FAILED 1 tests: ['test.CopterTests2b']

@amilcarlucas Can you give some direction?

roackb2 · 2020-11-20T13:23:39Z

And now it fails on cloning MAVProxy....
Any idea?

Installing MAVProxy
/tmp /__w/ardupilot/ardupilot
+ for t in $CI_BUILD_TARGET
+ '[' sitltest-tracker == sitltest-heli ']'
+ '[' sitltest-tracker == sitltest-copter-tests1 ']'
+ '[' sitltest-tracker == sitltest-copter-tests1a ']'
+ '[' sitltest-tracker == sitltest-copter-tests1b ']'
+ '[' sitltest-tracker == sitltest-copter-tests1c ']'
+ '[' sitltest-tracker == sitltest-copter-tests1d ']'
+ '[' sitltest-tracker == sitltest-copter-tests1e ']'
+ '[' sitltest-tracker == sitltest-copter-tests2 ']'
+ '[' sitltest-tracker == sitltest-copter-tests2a ']'
+ '[' sitltest-tracker == sitltest-copter-tests2b ']'
+ '[' sitltest-tracker == sitltest-can ']'
+ '[' sitltest-tracker == sitltest-plane ']'
+ '[' sitltest-tracker == sitltest-quadplane ']'
+ '[' sitltest-tracker == sitltest-rover ']'
+ '[' sitltest-tracker == sitltest-tracker ']'
+ run_autotest Tracker build.Tracker test.Tracker
+ NAME=Tracker
+ BVEHICLE=build.Tracker
+ RVEHICLE=test.Tracker
+ cat /proc/cpuinfo
+ '[' 0 -eq 0 ']'
+ echo 'Installing MAVProxy'
+ pushd /tmp
+ git clone --recursive https://github.com/ardupilot/MAVProxy
Cloning into 'MAVProxy'...
error: RPC failed; curl 56 GnuTLS recv error (-54): Error in the pull function.
fatal: the remote end hung up unexpectedly
fatal: early EOF
fatal: index-pack failed

roackb2 · 2020-11-20T14:10:45Z

@amilcarlucas finally it passed all the checks lolllllll
This is my first time doing a PR, but I'm wondering that is an nondeterministic CI/CD pipeline a common phenomenon?

amilcarlucas · 2020-11-20T14:30:25Z

Nice. yes, sometimes the CI is non-deterministic.

roackb2 · 2020-12-05T10:37:44Z

@amilcarlucas I've tested this feature with a copter for many times. It is really very helpful for my current project on hand, and I believe it would also be needed for many others studying indoor flight. To support those who might need this, I listed some information as following:

Here's the video
This is the flight plan
And this is the parameters I use
Hardware: Pixhawk 2.4
Vision: Intel Realsense T265
Companion Computer: Raspberry Pi 3B+
script running on companion computer to feed position data to flight control is here

amilcarlucas · 2020-12-05T14:27:35Z

ArduCopter/mode_auto.cpp

-        // failure to set destination can only be because of missing terrain data
-        copter.failsafe_terrain_on_event();
-        return;
+	// set waypoint controller target


The indentation looks messed up

amilcarlucas · 2020-12-05T14:29:27Z

libraries/AP_Mission/AP_Mission.cpp

+            packet.x = cmd.content.location.lat;
+            packet.y = cmd.content.location.lng;
+
+            packet.z = cmd.content.location.alt / 100.0f;   // cmd alt in cm to m


Please use a multiplication instead of a division here.

roackb2 · 2020-12-06T18:59:02Z

@amilcarlucas Thanks for your quick response. I got some errors consistently while running tests multiple times (listed below).

I dig into the code of test cases, looks like the test failed because of the difference in altitude of mission result. The code states that max_allowed_delta is 0.000009 for counter 4 ~ 10 in QGC WPL 110 file. The error points out that counter 10, which represents altitude, has difference of 0.000976999999693362 (1085.900391 - 11085.899414), which exceeds the max allowed delta. However, no related error shows up in last commit, and the only change in this PR that relates to altitude is to change the sign, since local frame is NED. The difference in this commit compared to prior one in AP_Mission.cpp is to change division to multiplication as you suggested. May there be any reason that this arithmetic change will result in the difference in mission execution? What kind of action should I take? I'm also curious about that do real vehicle driving or flying could achieve such kind of altitude accuracy?
Thanks.

AT-0142.2: Stopping SITL
AT-0143.3: Failing tests:
  DriveMission (Drive Mission balancebot1.txt) (Files have different (float) content: (/vagrant/Tools/autotest/ArduRover_Tests/DriveMission/balancebot1.txt) and (test.Rover/logs/2020-12-06/flight2/way.txt) (2	03	16	0.000000	0.000000	0.000000	0.000000	40.071083	-105.230026	11085.900391	1 vs 2	0	3	16	0.000000	0.000000	0.000000	0.000000	40.071083	-105.230026	11085.899414	1) (11085.900391 vs 11085.899414) (0.0009770000) (count=10)) (see /tmp/buildlogs/Rover-DriveMission.txt)
>>>> FAILED STEP: test.BalanceBot at Sun Dec  6 18:13:27 2020
Failure Summary:
  test.Rover:
    DriveMission (Drive Mission rover1.txt) (Files have different (float) content: (/vagrant/Tools/autotest/ArduRover_Tests/DriveMission/rover1.txt) and (test.Rover/logs/2020-12-06/flight1/way.txt) (2	0	316	0.000000	0.000000	0.000000	0.000000	40.071289	-105.230057	11085.900391	1 vs 2	0	3	16	0.000000	0.000000	0.000000	0.000000	40.071289	-105.230057	11085.899414	1) (11085.900391 vs 11085.899414) (0.0009770000) (count=10)) (see /tmp/buildlogs/Rover-DriveMission.txt)
  test.BalanceBot:
    DriveMission (Drive Mission balancebot1.txt) (Files have different (float) content: (/vagrant/Tools/autotest/ArduRover_Tests/DriveMission/balancebot1.txt) and (test.Rover/logs/2020-12-06/flight2/way.txt) (2	03	16	0.000000	0.000000	0.000000	0.000000	40.071083	-105.230026	11085.900391	1 vs 2	0	3	16	0.000000	0.000000	0.000000	0.000000	40.071083	-105.230026	11085.899414	1) (11085.900391 vs 11085.899414) (0.0009770000) (count=10)) (see /tmp/buildlogs/Rover-DriveMission.txt)
FAILED 2 tests: ['test.Rover', 'test.BalanceBot']

amilcarlucas · 2021-01-05T12:46:10Z

Can you change back to division and re-test?

amilcarlucas · 2021-01-07T12:55:42Z

I did the fixes and created a new PR #16259

amilcarlucas · 2021-01-07T13:09:24Z

#16259 is cleaner (no temp file), rebased, squashed and passes all CI tests

roackb2 · 2021-01-07T16:36:02Z

@amilcarlucas ok thanks for your work

chobitsfan added 2 commits November 19, 2020 23:21

AP_Common: add flag to Location indicating it uses LOCAL_NED

352666c

AP_Mission: support LOCAL_NED in takeoff and wp

dd3d838

roackb2 mentioned this pull request Nov 20, 2020

Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #8748

Closed

amilcarlucas suggested changes Nov 20, 2020

View reviewed changes

roackb2 force-pushed the pr_mission_local_frame2 branch from a29e361 to 49c4e8a Compare November 20, 2020 12:54

Copter: support LOCAL_NED in auto mode

e2aea23

roackb2 force-pushed the pr_mission_local_frame2 branch from 49c4e8a to e2aea23 Compare November 20, 2020 13:36

amilcarlucas suggested changes Dec 5, 2020

View reviewed changes

roackb2 force-pushed the pr_mission_local_frame2 branch from 2a298e7 to 27d54ed Compare December 6, 2020 15:07

amilcarlucas requested review from peterbarker and rmackay9 December 10, 2020 12:31

amilcarlucas mentioned this pull request Jan 7, 2021

Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #16259

Closed

Fix indentation

8eabaf3

roackb2 force-pushed the pr_mission_local_frame2 branch from 27d54ed to 8eabaf3 Compare January 7, 2021 12:48

amilcarlucas closed this Jan 7, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #8748 (rebased) #15854

Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #8748 (rebased) #15854

roackb2 commented Nov 20, 2020

amilcarlucas left a comment

roackb2 commented Nov 20, 2020

roackb2 commented Nov 20, 2020

roackb2 commented Nov 20, 2020

roackb2 commented Nov 20, 2020

amilcarlucas commented Nov 20, 2020

roackb2 commented Dec 5, 2020 •

edited

Loading

amilcarlucas Dec 5, 2020

amilcarlucas Dec 5, 2020

roackb2 commented Dec 6, 2020

amilcarlucas commented Jan 5, 2021

amilcarlucas commented Jan 7, 2021

amilcarlucas commented Jan 7, 2021

roackb2 commented Jan 7, 2021

Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #8748 (rebased) #15854

Copter: support MAV_FRAME_LOCAL_NED in MISSION_ITEM for auto mode #8748 (rebased) #15854

Conversation

roackb2 commented Nov 20, 2020

amilcarlucas left a comment

Choose a reason for hiding this comment

roackb2 commented Nov 20, 2020

roackb2 commented Nov 20, 2020

roackb2 commented Nov 20, 2020

roackb2 commented Nov 20, 2020

amilcarlucas commented Nov 20, 2020

roackb2 commented Dec 5, 2020 • edited Loading

amilcarlucas Dec 5, 2020

Choose a reason for hiding this comment

amilcarlucas Dec 5, 2020

Choose a reason for hiding this comment

roackb2 commented Dec 6, 2020

amilcarlucas commented Jan 5, 2021

amilcarlucas commented Jan 7, 2021

amilcarlucas commented Jan 7, 2021

roackb2 commented Jan 7, 2021

roackb2 commented Dec 5, 2020 •

edited

Loading