0 bank angle for GPS failure on landing #13795

Kjkinney · 2019-12-27T20:53:18Z

Describe problem solved by this pull request
On landing currently if an airplane enters a gps failure state, the plane will enter a fixed bank loiter, beginning a circle while on the landing path. What this does is allows the plane to instead keep the wings level in a failure instead of spiraling away from the land point.

Describe your solution
This commit checks if the plane is landing or not, and in a failure state. If so the navigator publishes a 0 bank angle in order to keep the plane heading towards the land point. Otherwise fixed bank loiter behaves exactly the same.

Describe possible alternatives
I used an alternate way of determining if the plane was landing within navigator_main. This version moves everything into gpsfailure. There might be a redundant boolean within the if statement currently.

Test data / coverage
This was tested in SITL in multiple different stages of flight to confirm nothing was broken when creating this.

https://review.px4.io/plot_app?log=683a837c-86d6-4d17-9330-2a47a87cc89f

Additional context
We have been flying with a different version of this on our own Fixed Wings for 6 months

@Antiheavy

…hicle is on slope for landing and enters a gps failure mode, the vehicle will continue straight ahead with wings level. Otherwise it behaves normally.

dagar · 2019-12-28T16:10:48Z

src/modules/navigator/gpsfailure.cpp

+			//This determines if the fixedwing vehicle is landing or not, and if so it forces the wings level
+			//Otherwise the plane goes to the parameter set fixed bank loiter setting
+
+			if (stat->nav_state == vehicle_status_s::NAVIGATION_STATE_AUTO_LANDGPSFAIL


Checking the nav_state is redundant. https://github.com/PX4/Firmware/blob/89e1f478ac9f3863df4f5613cf111222ae9c7749/src/modules/navigator/navigator_main.cpp#L617

dagar · 2019-12-28T16:11:12Z

src/modules/navigator/gpsfailure.cpp

+
+			if (stat->nav_state == vehicle_status_s::NAVIGATION_STATE_AUTO_LANDGPSFAIL
+			    && pos_sp_triplet->current.type == position_setpoint_s::SETPOINT_TYPE_LAND) {
+				att_sp.roll_body = 0;


Suggested change

att_sp.roll_body = 0;

att_sp.roll_body = 0.0f;

dagar · 2019-12-28T16:12:06Z

src/modules/navigator/gpsfailure.cpp

+			//Otherwise the plane goes to the parameter set fixed bank loiter setting
+
+			if (stat->nav_state == vehicle_status_s::NAVIGATION_STATE_AUTO_LANDGPSFAIL
+			    && pos_sp_triplet->current.type == position_setpoint_s::SETPOINT_TYPE_LAND) {


To be paranoid, also check that it's valid

Suggested change

&& pos_sp_triplet->current.type == position_setpoint_s::SETPOINT_TYPE_LAND) {

&& pos_sp_triplet->current.t.valid && (pos_sp_triplet->current.type == position_setpoint_s::SETPOINT_TYPE_LAND)) {

Kjkinney · 2020-01-02T16:02:23Z

Adding the way-point validity check to this breaks it currently. When it enters the fail safe mode, the current waypoint becomes invalid leading to it still using the fixed-bank loiter setting.

Kjkinney · 2020-01-02T16:13:52Z

Here is a log with the waypoint validity check: https://review.px4.io/plot_app?log=9f8842c3-bec7-4984-a6d8-03611e6d47a5

And here is a log without it: https://review.px4.io/plot_app?log=505571cc-7a93-4201-9367-80aa66963dcc

dagar · 2020-01-03T16:00:58Z

Adding the way-point validity check to this breaks it currently. When it enters the fail safe mode, the current waypoint becomes invalid leading to it still using the fixed-bank loiter setting.

Ok we'll need to think of something else then. What if you do the check in GpsFailure::on_activation()? Then GpsFailure could store a flag for was_landing and check that in the GPSF_STATE_LOITER case.

Kjkinney · 2020-01-03T16:37:22Z

Could we use Navigator::on_mission_landing() in the GpsFailure::on_activation() to store the was landing flag?

And would we want to do it in the GpsFailure::on_inactive() rather than on_activation() so that we don't run into any invalid issues with the gps failure activating?

EDIT
Just did a quick test using was_landing = _navigator->on_mission_landing(); within GpsFailure::on_inactive() and then checking it in the GPSF_STATE_LOITER case. It is working consistently at the moment.

…g gps failure.

Jaeyoung-Lim · 2020-01-09T08:17:03Z

src/modules/navigator/gpsfailure.cpp

+			} else {
+				att_sp.roll_body = math::radians(_param_nav_gpsf_r.get());
+			}
+
 			att_sp.pitch_body = math::radians(_param_nav_gpsf_p.get());
 			att_sp.thrust_body[0] = _param_nav_gpsf_tr.get();


Does it still make sense to use the same gps failsafe throttle? This might result in a fly away if misconfigured

Good point! Should we also force throttle to 0 if already landing?

After thinking about it more I think we should also force throttle to 0. If the vehicle was already landing, the intent is to get on the ground. In this case the best course of action is to continue wings level and cut throttle to ensure the vehicle gets on the ground. @Kjkinney can you add this code and code comments explaining why?

Kjkinney · 2020-01-09T17:18:58Z

I added the changes above. Here is a SITL test log

https://review.px4.io/plot_app?log=aff020ae-2efd-49ad-a207-30879d642283

Antiheavy · 2020-01-09T23:34:50Z

This PR will close this issue once approved and merged. #10906

@dagar is anything else required?

julianoes

I agree with the functional change but not with the implementation.

You are essentially changing the state called GPSF_STATE_LOITER, however, the more appropriate way to implement this would be to add a state, e.g. GPSF_STATE_LAND.

This has the benefit that you can announce the action correctly in on_activation(). Right now, it will display "Global position failure: fixed bank loiter" even though it is about to land which can be confusing for the operator.

I would therefore add a state and then check in advance_gpsf() which state you should go to depending on the was_landing flag. Then you can implement the correct setpoint in the respective state in on_active().

Do you think that would make sense?

RomanBapst · 2020-01-15T10:44:34Z

I agree that it would make more sense to split the two cases.

stale · 2020-04-14T11:21:20Z

This issue has been automatically marked as stale because it has not had recent activity. Thank you for your contributions.

This changes the fixed-bank-loiter failure for GPS failure. If the ve…

bff9cdb

…hicle is on slope for landing and enters a gps failure mode, the vehicle will continue straight ahead with wings level. Otherwise it behaves normally.

Antiheavy added Admin: Enhancement (improvement) 💡 Feature: Failsafe Fixed Wing 🛩️ labels Dec 28, 2019

dagar reviewed Dec 28, 2019

View reviewed changes

weekly-digest bot mentioned this pull request Dec 29, 2019

Weekly Digest (22 December, 2019 - 29 December, 2019) #13804

Closed

Kjkinney and others added 2 commits December 30, 2019 10:59

Add suggested changes

b44f35d

Merge branch 'master' into pr-gpsfailure-noBankLanding

7f7fc54

Remove waypoint validity check

9e238e5

Add a check to determine if the airplane was landing prior to enterin…

118cfe8

…g gps failure.

Antiheavy requested a review from Jaeyoung-Lim January 8, 2020 20:35

Jaeyoung-Lim reviewed Jan 9, 2020

View reviewed changes

Add the suggested 0 throttle in order to prevent flyaways.

c642f45

Antiheavy previously approved these changes Jan 9, 2020

View reviewed changes

Merge branch 'master' into pr-gpsfailure-noBankLanding

33b7409

julianoes requested changes Jan 15, 2020

View reviewed changes

stale bot added the stale label Apr 14, 2020

weekly-digest bot mentioned this pull request Apr 19, 2020

Weekly Digest (12 April, 2020 - 19 April, 2020) #14700

Closed

dagar dismissed Antiheavy’s stale review via 33b7409 June 7, 2023 16:07

stale bot removed the stale label Jun 7, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

0 bank angle for GPS failure on landing #13795

0 bank angle for GPS failure on landing #13795

Kjkinney commented Dec 27, 2019

dagar Dec 28, 2019

dagar Dec 28, 2019

dagar Dec 28, 2019

Kjkinney commented Jan 2, 2020

Kjkinney commented Jan 2, 2020 •

edited

Loading

dagar commented Jan 3, 2020

Kjkinney commented Jan 3, 2020 •

edited

Loading

Jaeyoung-Lim Jan 9, 2020

Antiheavy Jan 9, 2020

Antiheavy Jan 9, 2020

Kjkinney commented Jan 9, 2020

Antiheavy commented Jan 9, 2020

julianoes left a comment

RomanBapst commented Jan 15, 2020

stale bot commented Apr 14, 2020

	&& pos_sp_triplet->current.type == position_setpoint_s::SETPOINT_TYPE_LAND) {
	&& pos_sp_triplet->current.t.valid && (pos_sp_triplet->current.type == position_setpoint_s::SETPOINT_TYPE_LAND)) {

0 bank angle for GPS failure on landing #13795

Are you sure you want to change the base?

0 bank angle for GPS failure on landing #13795

Conversation

Kjkinney commented Dec 27, 2019

dagar Dec 28, 2019

Choose a reason for hiding this comment

dagar Dec 28, 2019

Choose a reason for hiding this comment

dagar Dec 28, 2019

Choose a reason for hiding this comment

Kjkinney commented Jan 2, 2020

Kjkinney commented Jan 2, 2020 • edited Loading

dagar commented Jan 3, 2020

Kjkinney commented Jan 3, 2020 • edited Loading

Jaeyoung-Lim Jan 9, 2020

Choose a reason for hiding this comment

Antiheavy Jan 9, 2020

Choose a reason for hiding this comment

Antiheavy Jan 9, 2020

Choose a reason for hiding this comment

Kjkinney commented Jan 9, 2020

Antiheavy commented Jan 9, 2020

julianoes left a comment

Choose a reason for hiding this comment

RomanBapst commented Jan 15, 2020

stale bot commented Apr 14, 2020

Kjkinney commented Jan 2, 2020 •

edited

Loading

Kjkinney commented Jan 3, 2020 •

edited

Loading