- Although the discussion is still ongoing: https://github.com/PX4/PX4-Autopilot/pull/20082#discussion_r962169411
- The original code would cause quick saturation in the steering input,
as it doesn't even have a conecpt of time
- So using direct mapping for now, but we can revert this back and find
the correct solution
- There were small points that were missed from the original PR: https://github.com/PX4/PX4-Autopilot/pull/20082
- This was tested in sync with the QGC PR, and was updated accordingly: https://github.com/mavlink/qgroundcontrol/pull/10483
- This modifies some of the parameters unit / range / name /
definitions, as some didn't make sense (e.g. GND_SPEED_MAX being used as
an output limiter for the velocity PID controller? the unit doens't
match)
- Also added clear unit documentation & function documentation to make
sure we don't mess up the radians / degrees, etc.
RoverPosControl: Multiply Roll command via Max Rate Parameter
With openjdk 17.0.5 2022-10-18, jmavsim fails to start due to missing
exports. This was previously a warning and turned into an error now.
Caught AppContextInfo(Bug 1004) InaccessibleObjectException: Unable to make public static sun.awt.AppContext sun.awt.AppContext.getAppContext() accessible: module java.desktop does not "exports sun.awt" to unnamed module @2c767a52 on thread J3D-Renderer-1
[0]: java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)
[1]: java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)
[2]: java.base/java.lang.reflect.Method.checkCanSetAccessible(Method.java:199)
[3]: java.base/java.lang.reflect.Method.setAccessible(Method.java:193)
[4]: com.jogamp.nativewindow.awt.AppContextInfo$1$1.run(AppContextInfo.java:40)
[5]: com.jogamp.common.util.UnsafeUtil.doWithoutIllegalAccessLogger(UnsafeUtil.java:219)
[6]: com.jogamp.nativewindow.awt.AppContextInfo$1.run(AppContextInfo.java:34)
[7]: java.base/java.security.AccessController.doPrivileged(AccessController.java:318)
[8]: com.jogamp.nativewindow.awt.AppContextInfo.<clinit>(AppContextInfo.java:31)
[9]: com.jogamp.nativewindow.awt.JAWTWindow.<init>(JAWTWindow.java:128)
[10]: jogamp.nativewindow.jawt.x11.X11JAWTWindow.<init>(X11JAWTWindow.java:60)
[11]: java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
[12]: java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:77)
[13]: java.base/jdk.internal.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
[14]: java.base/java.lang.reflect.Constructor.newInstanceWithCaller(Constructor.java:499)
[15]: java.base/java.lang.reflect.Constructor.newInstance(Constructor.java:480)
[16]: jogamp.nativewindow.NativeWindowFactoryImpl.getAWTNativeWindow(NativeWindowFactoryImpl.java:105)
[17]: jogamp.nativewindow.NativeWindowFactoryImpl.getNativeWindowImpl(NativeWindowFactoryImpl.java:66)
[18]: com.jogamp.nativewindow.NativeWindowFactory.getNativeWindow(NativeWindowFactory.java:654)
[19]: javax.media.j3d.JoglPipeline$QueryCanvas.addNotify(JoglPipeline.java:8604)
[20]: java.desktop/java.awt.Container.addNotify(Container.java:2804)
[21]: java.desktop/java.awt.Window.addNotify(Window.java:791)
[22]: java.desktop/java.awt.Frame.addNotify(Frame.java:495)
[23]: java.desktop/java.awt.Window.show(Window.java:1053)
[24]: java.desktop/java.awt.Component.show(Component.java:1728)
[25]: java.desktop/java.awt.Component.setVisible(Component.java:1675)
[26]: java.desktop/java.awt.Window.setVisible(Window.java:1036)
[27]: javax.media.j3d.JoglPipeline.getBestConfiguration(JoglPipeline.java:8379)
[28]: javax.media.j3d.Renderer.doWork(Renderer.java:496)
[29]: javax.media.j3d.J3dThread.run(J3dThread.java:271)
Caught AppContextInfo(Bug 1004) IllegalAccessException: class com.jogamp.nativewindow.awt.AppContextInfo cannot access class sun.awt.AppContext (in module java.desktop) because module java.desktop does not export sun.awt to unnamed module @2c767a52 on thread J3D-Renderer-1
[0]: java.base/jdk.internal.reflect.Reflection.newIllegalAccessException(Reflection.java:392)
[1]: java.base/java.lang.reflect.AccessibleObject.checkAccess(AccessibleObject.java:674)
[2]: java.base/java.lang.reflect.Method.invoke(Method.java:560)
[3]: com.jogamp.nativewindow.awt.AppContextInfo.fetchAppContext(AppContextInfo.java:191)
[4]: com.jogamp.nativewindow.awt.AppContextInfo.update(AppContextInfo.java:135)
[5]: com.jogamp.nativewindow.awt.AppContextInfo.<init>(AppContextInfo.java:50)
[6]: com.jogamp.nativewindow.awt.JAWTWindow.<init>(JAWTWindow.java:128)
[7]: jogamp.nativewindow.jawt.x11.X11JAWTWindow.<init>(X11JAWTWindow.java:60)
[8]: java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
[9]: java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:77)
[10]: java.base/jdk.internal.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
[11]: java.base/java.lang.reflect.Constructor.newInstanceWithCaller(Constructor.java:499)
[12]: java.base/java.lang.reflect.Constructor.newInstance(Constructor.java:480)
[13]: jogamp.nativewindow.NativeWindowFactoryImpl.getAWTNativeWindow(NativeWindowFactoryImpl.java:105)
[14]: jogamp.nativewindow.NativeWindowFactoryImpl.getNativeWindowImpl(NativeWindowFactoryImpl.java:66)
[15]: com.jogamp.nativewindow.NativeWindowFactory.getNativeWindow(NativeWindowFactory.java:654)
[16]: javax.media.j3d.JoglPipeline$QueryCanvas.addNotify(JoglPipeline.java:8604)
[17]: java.desktop/java.awt.Container.addNotify(Container.java:2804)
[18]: java.desktop/java.awt.Window.addNotify(Window.java:791)
[19]: java.desktop/java.awt.Frame.addNotify(Frame.java:495)
[20]: java.desktop/java.awt.Window.show(Window.java:1053)
[21]: java.desktop/java.awt.Component.show(Component.java:1728)
[22]: java.desktop/java.awt.Component.setVisible(Component.java:1675)
[23]: java.desktop/java.awt.Window.setVisible(Window.java:1036)
[24]: javax.media.j3d.JoglPipeline.getBestConfiguration(JoglPipeline.java:8379)
[25]: javax.media.j3d.Renderer.doWork(Renderer.java:496)
[26]: javax.media.j3d.J3dThread.run(J3dThread.java:271)
Exception in thread "main" java.lang.reflect.InvocationTargetException
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:568)
at org.eclipse.jdt.internal.jarinjarloader.JarRsrcLoader.main(JarRsrcLoader.java:61)
Caused by: java.lang.RuntimeException: java.lang.IllegalAccessException: class javax.media.j3d.JoglPipeline cannot access class sun.awt.X11GraphicsDevice (in module java.desktop) because module java.desktop does not export sun.awt to unnamed module @2c767a52
at javax.media.j3d.JoglPipeline.getScreen(JoglPipeline.java:8553)
at javax.media.j3d.Screen3D.<init>(Screen3D.java:354)
at javax.media.j3d.Canvas3D.<init>(Canvas3D.java:1124)
at javax.media.j3d.Canvas3D.<init>(Canvas3D.java:1026)
at javax.media.j3d.Canvas3D.<init>(Canvas3D.java:990)
at me.drton.jmavsim.Visualizer3D$CustomCanvas3D.<init>(Visualizer3D.java:909)
at me.drton.jmavsim.Visualizer3D.<init>(Visualizer3D.java:196)
at me.drton.jmavsim.Simulator.<init>(Simulator.java:193)
at me.drton.jmavsim.Simulator.main(Simulator.java:944)
... 5 more
Caused by: java.lang.IllegalAccessException: class javax.media.j3d.JoglPipeline cannot access class sun.awt.X11GraphicsDevice (in module java.desktop) because module java.desktop does not export sun.awt to unnamed module @2c767a52
at java.base/jdk.internal.reflect.Reflection.newIllegalAccessException(Reflection.java:392)
at java.base/java.lang.reflect.AccessibleObject.checkAccess(AccessibleObject.java:674)
at java.base/java.lang.reflect.Method.invoke(Method.java:560)
at javax.media.j3d.JoglPipeline.getScreen(JoglPipeline.java:8551)
... 13 more
* VTOL: take home_position into account for ground clearance
For approximating distance to ground in case there is no valid distance sensor
data, subtract the home position altitude from the local position altitude.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Co-authored-by: Matthias Grob <maetugr@gmail.com>
Instead of always starting with instance 0 (potentially an internal
mag), first take the current sensor as reference to compare the other
ones against it.
The issue is that otherwise we can end up in a
situation where a switch occurs because the currently used sensor isn't
much better than a sensor with a lower index: because the selected one
isn't much better, we cannot "fail-over" to it and get stuck on another
sensor, triggering a fail-over.
- WelfordMeanVector now computes covariance
- use Kahan summation for Welford mean (but continue using float32 for actual mean, etc)
- WelfordMean and WelfordMeanVector handle initial value and count roll over
- Welford mean count rollover at 16 bit max to prevent numerical issues and shift weight to newer samples
- sensors/vehicle_imu: update Welford mean usage (now simplified with resets removed)
- fix vehicle_imu_status accel var, now properly rotated with full covariance matrix
- gyro_calibration: update Welford mean usage
This is fully backwards compatible: If the throttle trim is set to
the minimum then it's the legacy calibration and gets
interpreted such that there is no trim and behavior remains as before.
If the trim is set to a different value than the minimum then it gets
used like with all other channels which was unsupported before.
In review it was requested to have a different name for
manual_control_setpoint.z because of the adjusted range.
I started to investigate what naming is most intuitive and found
that most people recognize the stick axes as roll, pitch, yaw, throttle.
It comes at no surprise because other autopilots
and APIs seem to share this convention.
While changing the code I realized that even within the code base
the axes are usually assigned to a variable with that name or
have comments next to the assignment clarifying the axes
using these names.
To be consistent with all other axes of stick input and avoid future
rescaling confusion.
Note: for the MAVLink message 69 MANUAL_CONTROL it's using the full range
according to the message specs now [-1000,1000].
Add ARK_FMU_V6X to RCS netman
Remove arkv6x rc single wire
Fix arkv6x mtd
arkv6x bootloader init all pins to prevent power cycling peripherals on boot
arkv6x don't power cycle sd card on boot
arkv6x add UART4 Telem 4
Also applies to Loiters that are started due to the previous mode being over (Takeoff,
VTOL_Takeoff, Mission).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
landing airspeed was previously defined by a scale factor multiplied by minimum airspeed. this commit changes this parameter to an explicit speed, and when unspecified, defaults to the minimum airspeed
Use kClearanceAltitudeBuffer for it, which is also used to ensure that during takeoff an
altitude setpoint above the clearance altitdue is set.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
MIS_LTRMIN_ALt was used to limit the go-to altitude of a LOITER_TO_ALT (not the exit altitude,
but the altitude that the vehicle went to to fly to WP), and during landing abort to climb to
at least this altitude. The min entry altitude of LOITER_TO_ALT I remove with this commit, while
for the min alt during abort I added the new parameter MIS_LND_ABRT_ALT.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
previously, the next position setpoint in the triplet was left unchanged during an abort, which meant that the abort loiter current point combined with land next point triggered the early landing config logic. this commit is only a hack to make things work temporarily.. this needs to be handled better.
- explicitly defined takeoff airspeed setpoint
- dont use climbout mode
- allow max climb on takeoff
- dont handle post clearance altitude case (navigator will switch anyway)
Completely detach the steering wheel logic from the yaw controller (beside using the
same manual stick input in a manual flight mode).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- also put flaring internal states into a struct to organize a bit
- one concern with blending the throttle setpoint like this with the flare time param is that folding prop belly landing airframes may want to have a separate param for shorter throttle kill and still use the flare time ramps for everything else
abruptly changing to a heading setpoint on flare can cause the aircraft to roll and deviate from the runway, this commit
- maintains path following control during the flare not to disrupt the tracking just before touchdown
- (unfortunately for crosswind landing) removes the body axis alignment for runway bearing - this is a compromise
to achieve both runway bearing body axis alignment AND a specific touchdown point, either
1. the wind would need to be considered, and an appropriate diagonal approach (obstructions allowing) defined to the runway
2. slip control added, keeping path following outputs only commanding roll (controlling airspeed vector) and using yaw-rate command (only actuated by e.g. rudder) to align body axis with the runway
- consolidate takeoff rotation transition times for pitch constraints and throttle setpoint with a single param
- consolidate pitch takeoff constraint parameters (remove rwto_max_pitch, use nominal max)
- input correct units to rwto pitch constraint getters
- encapsulate absolute time interpolator method for transitions
- start runway ops from idle throttle
TECS climbout mode was used for takeoff climbout, which puts throttle to full and does not regulate a specific airspeed.
This commit sets the desired takeoff airspeed explicitly and allows max climb rate to track the ascent.
previously a scale factor param on min airspeed was used to define the climbout airspeed for runway takeoff
additionally, the rotation speed was defined by another hardcoded scale on top of the previously scaled min airspeed
this commit explicitly defines a takeoff speed and rotation speed for runway takeoff in params, with option to disable
Before, the logic to update disarm and rate values also triggered during
bootup on the px4io, because the output functions are only set in
updateSubscriptions().
Therefore change the check to prevent updating during the first cycle.
When using ignition SITL simulation with NO_LOCKSTEP, the SITL PX4 fails to update the IMU data from Ignition Gazebo.
The timestamp for the IMU data is taken from the ignition message:
- In LOCKSTEP mode the clock from the ignition simulation and the one from PX4 SITL are synchronized, hence everything works fine
- In NO_LOCKSTEP mode, those clocks are not synchronized anymore, so the timestamp for the IMU data should not be the one from Ignition but the current time in PX4 SITL when receiving the message.
The update function of InputMavlinkGimbalV2 returns UpdatedNotActive when receiving control commands from a system or component that doesn't match the primary control. This can happen if a component just sends commands without being in control or when transitioning to a different primary control.
If the input is marked as already_active, it will reset last_input_active which in turn resets the primary control in the next iteration. According to the MAVLink gimbal protocol v2, a component needs to send MAV_CMD_DO_GIMBAL_MANAGER_CONFIGURE to start controlling the gimbal and to remove control. However, with the current implementation there are several other commands that would reset the primary control.
With this PR, the primary control remains with the component that requested it last if updates from a not active component are received.
* Log position setpoint reference of npfg
This commit logs the local position setpoint reference when using NPFG
* Address review comments
This commit address review comments from @tstastny
Replace the existing check for the availability of a takeoff mission item with a combined
check for takeoff and landing item (or landing pattern). New param MIS_TKO_LAND_REQ
can be set to require only a takeoff, only a landing, both takeoff and landing, and
both or none. The latter is meant to be set if is e.g. deemed unsafe to start a flight
through a Takeoff WP without though defining a Landing - as then in case of a RTL the
vehicle doesn't follow a pre-defined path but instead only can do default RTL that especially
for FW and VTOL isn't always safe.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Previously the minimum airspeed setpoint was adjusted to the load_factor compensated
stall speed, which, when the stall speed was set without margin, gave the controller
no room for error (the vehicle would stall if the controller has even a small airspeed
error).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- Include board_config.h for BOARD_GET_EXTERNAL_LOCKOUT_STATE etc. macros
- Include fcntl.h for "open"
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
* LandDetectorMC: enforce that LNDMC_Z_VEL_MAX * 1.2 is below *MPC_LAND_CRWL/MPC_LAND_SPEED
Otherwise the _in_descend flag doesn't get set correctly during the last part
of the landing, where the descend speed is at MPC_LAND_CRWL or LAND_SPEED.
The _in_descend flag is set it the velocity setpoint is >1.1*LNDMC_Z_VEL_MAX.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Co-authored-by: Matthias Grob <maetugr@gmail.com>
* Made voxl2 apps processor and slpi dsp processor builds into separate board builds so that they can
more easily be configured independently.
* Removed board specific link library command from px4_config.cmake and moved it to a more generic
board specific solution that can be used by any board that needs custom link libraries.
* Removed redundant cmake command for Qurt
* Removed unused definition from Qurt cmake file
* Removed unnecessary QURT_LIB cmake function
* Reorganized the voxl2 build structure to avoid 4 level board directories.
* Reverted cmake files to remove 4 level board naming code
* Updated documentation
- in a lot of cases this won't really matter, although you can see it in logs (especially with things like vehicle_at_rest), but it's something we might as well do properly
Co-authored-by: Mathieu Bresciani <brescianimathieu@gmail.com>
Fixes a regression from 8bae4e5c0e, where
the orbit flight task wasn't an extra task (flight_tasks_to_add) anymore
and therefore the command handling wasn't generated.
There was a race condition that could cause several outcomes. The most severe
was that flight_mode_manager gets the command, switches to orbit and then
in the next iteration switches back because commander did not change
nav_state yet. When commander then switches, flight_mode_manager would still
be in the old mode.
This is prevented by storing the command (allowing it to arrive before or
after mode switch), and then apply it after the switch happens.
- vehicle_angular_velocity and vehicle_angular_acceleration are produced together from the same input data, consumed together, and share the the same metadata (timestamps)
- individually these topics each have 16 bytes of metadata (2 timestamps) for 12 bytes of data (x,y,z float32)
During the whole flight, if the difference between the yaw estimate from
EKF2 and the emergency estimator is large and that the GNSS velocity
fusion is failing continuously, immediately reset to the emergency yaw
estimate.
- with Ekf::initialiseFilter() resetting _time_last_hagl_fuse the terrain estimate was technically valid regardless of any range or flow data availability and as a result optical flow fusion is able to start and stay active
- only consider terrain estimate valid based on control status flags and recent fusion
The CameraFeedback module used only the vehicle attitude for the camera orientation so far. With this change, the gimbal_device_attitude_status is used to compute the global camera orientation when a gimbal is used.
In this case it did not cause any problems.
Fixes a compiler warning:
/src/modules/commander/HealthAndArmingChecks/HealthAndArmingChecks.cpp:39:21: error: member ‘HealthAndArmingChecks::_failsafe_flags’ is used uninitialized [-Werror=uninitialized]
39 | _reporter(_failsafe_flags)
| ^~~~~~~~~~~~~~~
* simulator_mavlink: Add basic vio failure injection
Signed-off-by: Paul Frivold <paul@kefrobotics.com>
* simulator_mavlink: Rm failure not supported warning
Failures commands are also handled in other files,
so warning here could be confusing.
Signed-off-by: Paul Frivold <paul@kefrobotics.com>
Signed-off-by: Paul Frivold <paul@kefrobotics.com>
previous change in logic to hold after mission clear also broke rtl, as non-mission takeoff still published a mission result which allowed entering the mission finished condition and always changing state to loiter (ignoring rtl). new logic only switches navigation states if mission is finish and the nav state is explicitly in takeoff state, or in mission state
The srcparser.py is specific to each use case (e.g. Airframes, Parameters, px4events, etc as in Tool/* folders).
Therefore it is confusing to have the px_process_airframes.py script handle concept of airframes under the generic name 'params'.
This improves readability and sets the baseground for implementing more specific vehicle type supports, as mentioned in https://github.com/PX4/PX4-user_guide/pull/1858#discussion_r876554728
once the rampup starts. The rampup requires a valid vertical velocity setpoint.
The corner case is:
- We are before takeoff and amending the setpoint to be 0,0,100 acceleration
in order to idle
- The rampup starts BUT the setpoint is not yet overwritten by the trajectory
setpoint topic
- The idle setpoint gets amended to not contain a feed-forward vertical
acceleration because the rampup is velocity based
- The result is a brief invalid 0,0,NAN acceleration setpoint
- That invalid setpoint gets overridden by a failsafe that holds zero velocity
- Zero velocity leads to applying ~hover thrust briefly
- buffer at least 2 samples for the IMU output predictor buffers
- dropping below 2 becomes problematic for the minimum observation
interval calculation and the vertical output buffer trapezoidal
integration
arm
- Previously, due to the way MAVSDK's `health_all_ok` was implemented,
vehicle often didn't have a valid global position estimate (although
function returned true), and it wouldn't arm, and the SITL would fail
- Also sometimes as vehicle didn't have manual control, it entered weird
states where it wasn't able to arm as well
- This adds a check to make sure vehicle is able to arm, directly from
the Health struct
- preflight mag bias estimate is in the vehicle body frame and removed from the raw mag data after it's rotated, calibrated, etc
- in flight mag bias (from ekf2) is in the vehicle body frame, but stored as a sensor frame offset immediately
* Enable motor controls for fixed wing mode in tailsitters
This commit enable motor controls in fixed wing mode for tailsitters
This is needed for enabling quad tailsitters
* VTOL: differential thrust in FW: adapt params to be generic for all axes
Until now only suppoted on yaw axis. Not to be supported also on Roll and Pitch.
- VT_FW_DIFTHR_EN: make bitmask for all three axes independently. First bit is Yaw,
sucht that existing meaning of VT_FW_DIFTHR_EN doesn't change.
- VT_FW_DIFTHR_SC: rename to VT_FW_DIFTHR_S_Y and add same params for roll (_R) and
pitch (_P).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
* Integrate differential control bits to three axis control
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
It's enough that the setpoints and the unallocated values are logged, from these
the allocated values can be calculated if required.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
It's very important that the in descend detection is always
at a strictly higher velocity than the vertical movement check.
This combination is basically used to check for vertical downwards
velocity tracking. Desired descend, no movement -> ground
If in descend threshold is lower than vertical movement it is
by definition even with perfect tracking the case that with
any velocity between the two thresholds there is no movement
even though a descend is commanded.
See first fix of this problem #7831e39b38ba96
I don't see where this is necessary. During takeoff, the maybe landed flag should
only get cleared once system is about to takeoff, and thus well after the spool up
is complete (for which the higher thresholds are meant in this case).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Don't consider these params for vertical speed threshold,
and instead increase the default for LNDMC_Z_VEL_MAX and
use solely that one. Makes the land detector outcome more
predictable and its interal logic simpler, while the new
default tuning is resulting in about the same vz threshold
as before.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
* Changed M_PI to M_PI_F in the matrix library since M_PI is non-standard.
* Added a new M_PI_PRECISE constant definition to px4_platform_common/defines.h to be
used in places when M_PI is desired but shouldn't be used because it is not C standard.
* Added the px4_platform_common/defines.h include to the matrix library math.hpp header to pull
in some non-standard M_PI constants and updated the test files to use those constants.
* Fixed PI constants in matrix helper test to prevent test failure
before there was a corner case where if in an auto mode that sets a reset command on the attitude setpoint message (e.g. auto takeoff), if the mode was then switched stabilized, this reset bool would never be changed back to false and the integrators would reset every cycle
* Removed dprintf from perf library since it is only ever used with fd=1 (STDOUT) so moved to PX4_INFO_RAW instead. This helps with some platforms (e.g. Qurt) which have some Posix support but not full Posix support.
This was previously required to reset the flight speed after a VTOL transition,
but is now no longer required as the DO_CHANGE_SPEED commands are handles directly
in the controllers.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
This fixes the issue where the init happended in the initializer list, at which point
the params were not yet initialized and thus resulted in random values for the init
values of _mission_item.loiter_radius and .acceptance_radius.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Refactor
- Require reboot for PD_GRIPPER_EN parameter change
- Define gripper ACTION_NONE for readability. This makes implicit assumption that -1 equals no-action commanded more explicit
- Tidy the scattered vcmd_ack struct handling cases into a single function
- Refactor to remove return in the middle of function: avoids future complications where a programmer may expect the logic at the end of the function to be executed, but isn't
Vehicle Command Handling
- Cancel the previous running vehicle command if we receive a different vehicle command
- Reject vehicle command if we get a same one that is getting executed
- Save the source system & component of currently running vehicle command
- Added note about new discovered edge case of having same entity sending different gripper commands consequently, where an unexpected ack result may be received
Gripper:
- Don't command gripper (via uORB `gripper` topic, which maps into an
actuator via Control Allocation) if we are already at the state we want
(e.g. Grabbed / Released) or in the intermediate state to the state we
want -> This prevents spamming on `gripper` topic
Payload Deliverer:
- Add read-once function for Gripper's released / grabbed state
- Send vehicle_command_ack for both release/grab actions.
TODO: target_system & target_component for the released/grabbed vcmd_ack
is incomplete, since we are not keeping track of the vehicle_command
that corresponds to this. This needs to be dealt with in the future, not
sure what the best solution it is for now.
Possible solutions:
- Queue-ing the vehicle command?
- Tying the gripper's action to specific vehicle command one-on-one, to make sure if we send multiple vehicle commands, we know
which command resulted in the action exactly?)
Only command Gripper grab when we are actually initializing gripper
- Previously, on every parameter update, gripper grab was being
commanded
- This commit narrows that scope to only when we are actually
initializing the gripper
Handle gripper de-initialization upon parameter change
- Also added some local state initialization code to init() function of
Gripper
- This will now make init / de-init more graceful & controlled compared
to before
- This hopefully then alerts the GCS that the command is getting
processed
- Referenced commander's `handle_command` function to implement this. As
it seems like GCS needs the acknowledgement of the command being
processed to execute such commands properly
- Also send FAILED command ack if we can't actuate the gripper
Fix wrong GRIPPER_ACTION conversion from floating point to int32_t
- Due to the MAVLink spec, we actually just convert enums into floating
point, so in PX4 we need to convert the float directly into integer as
well (although there can be precision issues on large numbers)
- This is a limitation in MAVLink spec, and should hopefully be
changed in MAVLink v2
Affects the states AUTO_ALTITUDE and AUTO_CLIMBRATE. Those modes should only be entered
if armed (as they are pure failsafe modes). Also allow though to enter them even if
the position setpoint(s) are invalid, as they are not needed.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Instead of checking the .valid flag of position_setpoint, check for ISFINITE() of lat, lon, alt
when pulling the position_setpoint triplet. This fixes problems where the .valid flag didn't
reflect the proper state of the setpoint (e.g. .valid was true, .lat though NAN)
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
the nuttx and posix specific options files since this option cannot be used with
the qurt platform. There are header files in the hexagon sdk that fail this check.
as Tools/update_px4_ros2_bridge.sh as been deleted
update_ros2_bridge, update_px4_ros_com and update_px4_msgs
are no more needed
Signed-off-by: Beniamino Pozzan <beniamino.pozzan@phd.unipd.it>
- vehicle_angular_velocity: ESC RPM notch filters minimize filter resets
- only allow one filter init per axis per cycle
- "park" ESC notch filters at min frequency instead of full disable
- relax timeout before a notch filter is disabled
- add new parameter IMU_GYRO_DNF_MIN for configuring the minimum notch filter frequency
- avoid the possibility of unsigned underflow from subtracting two HRT timestamps (uint64_t)
- most of these aren't problematic, but people tend to replicate the pattern, so it's better to be safe
- likely wasn't a problem when people were using hrt_absolute_time() in place, but if using an existing timestamp there's the possibility it's older than a more recent topic update
- update all msgs to be directly compatible with ROS2
- microdds_client improvements
- timesync
- reduced code size
- add to most default builds if we can afford it
- lots of other little changes
- purge fastrtps (I tried to save this multiple times, but kept hitting roadblocks)
- working towards keeping all height source (baro/ev/gnss/rng) handling as consistent as possible, possibly refactoring these out into separate classes later
This fixes a bug where by accident the vtol_status was considered instead of the
vehicle_status, preventing it from running on planes.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
This commit fixes feedforward acceleration setpoints for fixedwing offboard position control.
Previously when acceleration feedforward inputs were sent, negative curvature accelerations were not being computed properly
ekf2: merge mag 3d innov var, Hx and Kx computation to reduce flash
Slightly less code produced, almost no performance change
ekf2_mag3D: do not pre-compute Kalman gains
The vector of Kalaman gains is not too expensive to compute using
matrix-vector multiplication. Pre-generating it using CSE takes a lot of
flash space for little benefit.
- this is a precaution to eliminate the possibility of getting stuck in
a loop trying to keep up with a high rate publication that could be
coming from a higher priority task
- in the worst case scenario printing the error message can take longer than the next gyro publication, so combined with an infinite loop you could get stuck here
The previous default for yaw and roll was 0.2, which is very low, and for wheel was
1, which is very high. A value of 0.4 makes sense to me for all axes.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
I guess this was a left over from times where the controller had a completely different structure,
a value of 50 is well outside of reasonable range (50x the default).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
The previous max of 10 for the FF gains seems a bit very high for me, well outside
of reasonable reange.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- Fixed clearing arming word if flightmode unknown
- Added power and cell voltage elements
- New OSD layout
- Fixed home direction/distance are now correct
Co-authored-by: Chris Seto <chris1seto@gmail.com>
Moves the arming checks before the command handling.
This reduces the chance of race conditions. However it does not prevent
them! The SDK will need to check when offboard is ready to run/arm to fix
this.
Specifically this is for sitl offboard tests, where offboard_control_mode
is updated and immediately after a mode switch into offboard is commanded.
- trajectory_setpoint and vehicle_local_position_setpoint used to be
the same data type
- we got extremely lucky here that this didn't cause any issues due to
all the fields still aligning
Starting the fake position as soon as all the aiding sources stop makes
the local position immediately invalid while we could continue to navigate
for a couple of second with inertial dead-reckoning
- vehicle_imu_status can publish immediately on any measured sample
rate change or sensor error increment, but the windowed mean/variance
shouldn't necessarily reset that often
This is to prevent that that data from non-recent publications of the virtual
attitude setpoints are used in the transition code.
It removes the previous implementation where the update_transition_state()
method was only run when both mc and fw virtual att sp where recent, independetly
of whether both are used or not.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- for accel/gyro/mag estimated bias only consider them stable (valid
for calibration updates) if the value isn't changing (10% of limit) over
the validity period
* Changed sitl_gazebo to the advanced liftdrag plugin fork.
* Added advanced plane and its associated parameters to PX4. Also tweaked one of the plugins to fix a compilation error (upcasting from float to double).
* Switched gazebo back to main branch, to avoid merge conflicts.
* Change simulator bridge back to what it was in main branch
* Changed sitl_gazebo to match the PX4 main branch's (commit hash b968405)
* Changed SimulatorIgnitionBridge to match most recent main.
* Removed exit after dsp signature generation.
* First full test suite
* Cleaned up the muorb tests
* Improved VOXL 2 self test trigger in startup file
* Removed unneeded include file
This removes the odd px4_i2c_bus_external override which was confusing
me and lead to odd and inconsistent results.
The function is now only available with an int as the argument.
- Previously it was hard to check which state the gripper was in
(GRAB/GRABBED/ etc), not even whether it was correctly initialized, this
commit adds that functionality
- When the gripper was uninitialized, the error message popped up
everytime the module had 'any' vehicle command received. But we should
be spamming only when we do actually get 'DO_GRIPPER' command received,
this commit adds that feature
- Also converted error messages for CLI when gripper test functinality
fails, so that user would get a feedback (originally, the PX4_DEBUG was
of course not outputting everything)
Also containst a mini refactor by removing the unnecessary function
get_default_altitude_acceptance_radius().
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
This new check is inteded to trigger if there is no data variation published by
the airspeed driver for the first 10s after the first data is published.
This is to capture malfunctioning sensors/drivers that do publish a value, this
value though does not come from real sensor measurements.
Previously this was captured by the data stuck check, but it has shown that
some drivers can publish a stuck value without being actually malfunctioning
(e.g. when the airspeed is outside of their measurement range). Checking for
any data variation is the more conservative check that still catches the above
described failure case.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
When a mission starts capturing images using the
MAV_CMD_IMAGE_START_CAPTURE command, it should also stop it again, in
case it is stopped early, e.g. with RTL.
Issue addressed: ROS2 is built from source and
no system-wide version of fastrtps is installed
Signed-off-by: Beniamino Pozzan <beniamino.pozzan@phd.unipd.it>
Such that the loiter (orbit) status can be displayed on groundstation
and the WP is accepted once within loiter radius
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- To check stack smashing (buffer overflow)
- To check if the buffer without enough length gets appropriate values
filled
- To check if the format has expected length
- Previously, not having a proper boundary check caused overflows in the
buffer (wrong memory access)
- Moreoever, negative size values getting introduced to snprintf &
strncat were also being truncated to unsigned value, hence causing
overflow, so index check needed to be added before both functions
- Fixed typo in board_common.h
If the current yaw setpoint is valid we should use it to make the transition in this direction.
For a VTOL_TAKEOFF the yaw_setpoint is used to specify the transition direction (the
vehicle is aligned towards it and then transition is started). The current yaw can be
a bit off (e.g. because MIS_YAW_ERR is large), and it is better to track the actual setpoint.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- for convenience merge px4_sitl_ign into px4_sitl_default, but allow simulator_ignition_bridge to quietly skip inclusion if ignition-transport isn't available
- simulator_ignition_bridge only try setting the system clock in
lockstep builds
- this simplifies usage and CI system dependencies
- rcS parameter backup try to directly restore param (FRAM) from backup (in case SD card is removed before successful export)
- rcS parameter backup logging rearrange to capture more logging output (param_import_fail.txt)
- posix rcS try to keep param backup and restore roughly in sync with NuttX rcS
- tinybson fix debug printf format
- param_export_internal ensure file descriptor positioned at 0 (precaution)
This state is added to give extra time between FIFO flush command. Some icm42688p IMUs need more time between config -> FIFO reset -> FIFO read. More about the issue #20181
Please use [PX4 Discuss](http://discuss.px4.io/) or [Slack](http://slack.px4.io/) to align on pull requests if necessary. You can then open draft pull requests to get early feedback.
<!--
## Describe problem solved by this pull request
A clear and concise description of the problem this proposed change will solve. Or, what it will improve.
E.g. For this use case I ran into...
Thank you for your contribution!
## Describe your solution
A clear and concise description of what you have implemented.
Get early feedback through
- Dronecode Discord: https://discord.gg/dronecode
- PX4 Discuss: http://discuss.px4.io/
- opening a draft pr and sharing the link
## Describe possible alternatives
A clear and concise description of alternative solutions or features you've considered.
-->
## Test data / coverage
How was it tested? What cases were covered? Logs uploaded to https://review.px4.io/ and screenshots of the important plot parts.
This repository holds the [PX4](http://px4.io) flight control solution for drones, with the main applications located in the [src/modules](https://github.com/PX4/PX4-Autopilot/tree/main/src/modules) directory. It also contains the PX4 Drone Middleware Platform, which provides drivers and middleware to run drones.
@@ -75,7 +75,7 @@ The PX4 Dev Team syncs up on a [weekly dev call](https://docs.px4.io/main/en/con
See also [maintainers list](https://px4.io/community/maintainers/) (px4.io) and the [contributors list](https://github.com/PX4/PX4-Autopilot/graphs/contributors) (Github).
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Junwoo Hwang