Plane: fixed takeoff direction with no yaw source #28730
Conversation
in TAKEOFF mode with either very poor yaw source or no yaw source we need to use ground vector and wait for sufficient ground speed
tridge
force-pushed
the
pr-takeoff-direction-fix
branch
from
a5fc6b1
to
640f205
Compare
| eventually trigger | ||
| */ | ||
| if (!plane.throttle_suppressed && | ||
| groundspeed > plane.aparm.airspeed_min*0.3) { |
There was a problem hiding this comment.
So basically you will not allow the heading to "lock-in" until some groundspeed is gained, deeming the heading valid.
Sounds reasonable.
| if abs(bearing_from_home - start_loc.heading) > bearing_margin: | ||
| raise NotAchievedException(f"Did not takeoff in the right direction {bearing_from_home}") | ||
|
|
||
| self.fly_home_land_and_disarm() |
There was a problem hiding this comment.
Since the test is only about the takeoff, does it make sense to not wait for the landing, but end the test early here? This would alleviate the test server a bit.
There was a problem hiding this comment.
shouldn't the course lock based on GPS course be based on when the GPS speed is valid? and this is not really airspeeed dependent at all?
- in Auto takeoffs, its set 2s after exceeding 5m/s ground speed from GPS (which I think could be set immediately after exceeding this since modern GPS's have been accurate for years, unlike when this bit was written)
- in my pending Autoland Lua script PR , I use 5m/s with a modern cheap GPS (M8) and it works great capturing takeoff direction, even in very strong crosswinds in compassless hand launches
- your criteria is 2.7m/s in the default case anyway
I think that it should just be set to a magic number of X (3-5m/s?) since the quality of the GPS heading is not really related to anything to do with airspeed
|
Off-Topic Question - Does takeoff compensate for cross-wind ? Because until now I was convinced that the heading error is also present in auto takeoff mission item. |
|
No navigation other than heading lock |
ArduPlane/mode_takeoff.cpp
Outdated
| const Vector2f &groundspeed2d = ahrs.groundspeed_vector(); | ||
| const float direction = wrap_360(degrees(groundspeed2d.angle())); | ||
| const float groundspeed = groundspeed2d.length(); | ||
|
|
||
| // see if we will skip takeoff as already flying | ||
| if (plane.is_flying() && (millis() - plane.started_flying_ms > 10000U) && ahrs.groundspeed() > 3) { |
There was a problem hiding this comment.
| if (plane.is_flying() && (millis() - plane.started_flying_ms > 10000U) && ahrs.groundspeed() > 3) { | |
| if (plane.is_flying() && (millis() - plane.started_flying_ms > 10000U) && groundspeed > 3) { |
There was a problem hiding this comment.
Is there a chance that the ground speed check here will trigger before the new one on line 129. If your min airspeed is larger than 10m/s (3/0.3) and it takes longer than 10 seconds to get there then its will skip the full takeoff every time. Really this skip takeoff path should only be done once. Maybe it should be moved to the mode enter?
tridge
force-pushed
the
pr-takeoff-direction-fix
branch
from
640f205
to
6b73c7f
Compare
in TAKEOFF mode with either very poor yaw source or no yaw source we need to use ground vector and wait for sufficient ground speed
fixes #28681