* MAVSDK_Test_Runner: Place PX4 instance runner after Gazebo server runner
- This was a nasty bug where starting PX4 instance first, then starting
Gazebo server was causing PX4 instance' EKF to freak out, probably
because it doesn't like getting data a while after it is started
- Detailed breakdown is given here: https://github.com/PX4/PX4-Autopilot/issues/21229#issuecomment-1450761542
- This now guarantees that such edge case won't occur and MAVSDK test
will run as it should
* MAVSDK Test Runner: Retain comment within 79 character limit
- To pass flake8 python style check
* Update test/mavsdk_tests/mavsdk_test_runner.py
Co-authored-by: Julian Oes <julian@oes.ch>
---------
Co-authored-by: Julian Oes <julian@oes.ch>
Testing has shown that 10m is a bit too tight, most of all as this check is also run in not height-controlled
flight modes (eg Stabilized), and there 10m altitude loss during transitions can happen easily.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
This adds the Cube Orange+ to the list, and also changes the Hex naming
to CubePilot as that is how it is sold/marketed now.
Signed-off-by: Julian Oes <julian@oes.ch>
The attitude frame is wrong for tailsitters doing side slip fusion for wind estimation.
It doesn't take into account that the frames is 90deg tilted in FW flight.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
-set roll/pitch used for failure detection during transition to 0
-rotate estimated attitude 90° in FW flight
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
to enable proper automatic output in gtest unit tests
that compare two `Matrix`, `Vector` or two `Dual` objects.
Credits to @jwidauer for showing me this trick.
Maximum plaintext length was wrong, it is just k - 2 * hLen -2, where k is RSA key modulus length and hLen is hash length
Also minimum block that can be encrypted didn't make sense for RSA-OAEP, it is just 1 byte, the rest will be padded.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Bootloader needs to have a mechanism to de-initialize crypto, in case some HW accelerator
is being used. This adds the needed function for it
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Camera controls should not happen through the flight controller, and
the control allocation has no means of controlling the camera zoom.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- add unit tests for adsb conflict detection
- move adsb conflict detection to lib/adsb and adsb conflict class
- use containers/Array.hpp for buffer array
- expand fake_traffic
This adds a parser for Unicore sentences sent in-between NMEA sentences.
This brings support for the Unicore UM982 module which publishes heading
information based on the two antennas.
Signed-off-by: Julian Oes <julian@oes.ch>
If NuttX is built without support for SMPS it can brick the hardware.
Therefore, I suggest that we add this additional compile-time check.
Signed-off-by: Julian Oes <julian@oes.ch>
It turns out that when you rotate by 45 degrees, as required on the
CubeOrange+, then you can easily get into clipping because the vector
components are constrained after the rotation. In order to avoid that,
we have to avoid getting close to the int16 range and switch from 20 bit
resolution to 16bit resolution earlier.
Signed-off-by: Julian Oes <julian@oes.ch>
This switches from attitude_estimator_q to EKF2 which should now work
without mag when the params are set to SYS_HAS_MAG = 0 and
EKF2_IMU_CTRL = 7 to enable gravity fusion.
Signed-off-by: Julian Oes <julian@oes.ch>
PX4_SIM model need the simulator (gz_) prefix
Fix post debug task
Add x500_depth, rc_cessna, standard_vtol
Signed-off-by: Beniamino Pozzan <beniamino.pozzan@phd.unipd.it>
- Add the possibility in the parser to replace the defines made in the current file with their argument (includes are not supported)
- Add the possibility for the parser to parse int argument with bitwise shift operators
- remove deprecated actuator_controls[INDEX_FLAPS/SPOILERS/AIRBRAKES]
- use new topic normalized_unsigned_setpoint.msg (with instances flaps_setpoint
and spoilers_setpoint) to pass into control allocation
- remove flaps/spoiler related fields from attitude_setpoint topic
- CA: add possibility to map flaps/spoilers to any control surface
- move flaps/spoiler pitch trimming to CA (previously called DTRIM_FLAPS/SPOILER)
- move manual flaps/spoiler handling from rate to attitude controller
FW Position controller: change how negative switch readings are intepreted
for flaps/spoilers (considered negative as 0).
VTOL: Rework spoiler publishing in hover
- pushlish spoiler_setpoint.msg in the VTOL module if in hover
- also set spoilers to land configuration if in Descend mode
Allocation: add slew rate limit of 0.5 to flaps/spoilers configuration change
Instead of doing the flaps/spoilers slew rate limiting in the FW Position Controller
(which then only is applied in Auto flight), do it consistently over all flight
modes, so also for manual modes.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
- remove deprecated actuator_controls[INDEX_LANDING_GEAR]
- remove dead code in mc rate controller that used to prevent it from being retracted
on the ground (anyway had no effect as it only affected the actuator_control[LANDING_GEAR]
which wasn't sent to the control allocation)
- for VTOLs handle deployment/retraction of landing gear in AUTO as a MC (retract if
more than 2m above ground, deploy if WP is a landing WP), plus additionally when transition
flight task is called (ALTITUDE mode and higher)
- for FW in AUTO: add logic in FW Position Controller, depending on waypoint type mainly
- manual landing gear settings always come through (a manual command overrides a previous
auto command, and vice-versa)
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Tiltrotor_extra_controls also contains collective thrust beside collective tilt, as passing a 3D
thrust setpoint vector beside the tilt is not feasible.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
* Changed the method of checking and setting the server file lock on Posix to avoid conditions where the server can indicate that it is running but still hasn't finished it's initialization
- always publish esc_status
- when enabled via MODAL_IO_VLOG param, enable actuator debug output
- for modal_io commands, use ESC HW ID values instead of motor number for easier use
- publish esc_status message for command line commands
- Uncommented the code that fills in the cmdcount and power fields in the esc_status topic
---------
Co-authored-by: Travis Bottalico <travis@modalai.com>
- if timestamp (or timestamp_sample) unpopulated fill it with current hrt_absolute_time()
- if using provided timestamp don't allow it to exceed current hrt_absolute_time()
- Switching to the first order filter that was previously
only in FlightTaskManualAltitude.
- Moving the scaling of full stick deflection to
radians per second into the class.
* RTL: only do calculations in is-inactive if global position is recent
* RTL: refactor calcRtlTimeEstimate to only calc and not pub
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
When requesting a message from a stream that is not active we start the
stream with interval=0 and call the request method once. For all streams
this works fine except the gps_global_origin. For this one the request method
is actually overidden to throttle down the rate and not just send out the message.
This will cause this message to never being sent on request if the stream
is not active by default.
* Update README: Maintainers, Boards, Roadmap
- First step after the community coordination call from January 30th
* README: Fromat list & remove discontinued boards & add others
- Addressed comments
* README: Add Simulation, remove QGC
- Only leave the PX4 specific categories (QGC is not)
* Add Beniamino as ROS2 maintainer
* README: Add note that README is main source of truth for maintainers
- We need to have a source of truth, we can use Github README for that.
- vectornav library (libvnc) fixed for NuttX
- open serial port O_NONBLOCK (like __APPLE__)
- set serial port baud rate with cfsetspeed (like __APPLE__)
- vectornav backend thread increase stack and run at higher priority (SCHED_FIFO)
Block Device driver uses a buffer so we need to ensure data is written or read to the device and not to the buffer so we can be sure if the device works properly
* Cyphal: fix comparing floating-point issue
* Cyphal: fix setpoint serialization
* Cyphal: fix bug with wrong comparasion of param name and pub/sub name: remove prefix from UavcanPublisher::updateParam and UavcanDynamicPortSubscriber::updateParam and PublicationManager::updateDynamicPublications
* Cyphal: integrate UavcanEscController with PublicationManager, remove second instance of UavcanEscController from CyphalNode
* Cyphal: publish readiness with minimal frequency because according to UDRAL The drive shall enter STANDBY state automatically if the readiness subject is not updated for CONTROL_TIMEOUT
* Cyphal: increase setpoint publish rate from ~75 to 200 by removing PX4_INFO (it really significantly react on the the output rate) and changing the mixing output rate and the shedule interval
* Cyphal: restore prefix because we need it for uorb over uavcan/cyphal and add udral prefix for non uorb pub/sub
* Cyphal: fix DynamicPortSubscriber subscription: if it has multiple subscribers, it should call subscription only after updating of all port subscribers port identifiers
* Cyphal: fix SubscriptionManager: we should take care about prefix
* Cyphal: fix readiness for test motor mode
* [Cyphal] Fix dynamicsubscription, improve printinfo, enable MR-CANHUBK3 config
---------
Co-authored-by: Peter van der Perk <peter.vanderperk@nxp.com>
- pass new airspeed sample around when available
- can't completely eliminate _airspeed_sample_delayed until resetWind()
called from sideslip fusion is updated
- reduce boost priority to PX4_WQ_HP_BASE - 6
- add cli command 'trigger_watchdog' to manually trigger watchdog
- add perf counters when triggering watchdog
- reduce top measurement to 300ms
- restore priorities after 1.5s
There are precautions in case the SD card code itself has a busy-loop.
Enable automatic landing abort on timed out distance sensor reading also for
the circular landing. Do not enable the no-terrain timeout check, as, opposed
to the straight landing, we here don't know when to expect the distance sensor
to get valid.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
As we don't know the landing point altitude in non-mission landings, assume
the worst case (abort right before touchdown) and thus always climb
MIS_LND_ABRT_ALT on triggering an abort.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Add method for circular landing, that is used instead of the straight fixed-wing
landings in case the landing is not part of a mission landing.
Use straight landing if previous WP is valid, and the ciruclar otherwise.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
_land_start_index is used to to start the mission from this item index, and to
avoid to publish a triplet.current.type=IDLE, we need to fill it with the actual
position setpoint that the vehicle should go to at the start of a mission landing.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Prior commit added opening of /dev/null as 0, 1 and/or 2 file
descriptors, if they where not present. However, if the temporary
file descriptor used to open /dev/null matched the target file
descriptor, it would be immediately closed again. This commit fixes that,
and does not duplicate and close the temporary file descriptor if it is
already at the correct number.
- new FIFO_RESET state used to give the sensor more time after successful configuration before sampling begins
Co-authored-by: Mathieu Bresciani <brescianimathieu@gmail.com>
* rtl: remove unconditional transition to land after descent
This was a bug, as it renders the above code lines useless.
This would cause a undesired FW landing for VTOL vehicles if
RTL_LAND_DELAY is above 0.
* rtl: head to center after loiter in VTOL FW
To get the same behavior for RTL with and without loiter before land for
VTOL drones.
* rtl: always go to descend state after return
Previously, the state would change directly to land if in MR and
RTL_LAND_DELAY was 0.0, but we will still wish to descent to
RTL_DESCEND_ALT at descent speed, instead of using landing speeds.
* rtl: mark head to center state as part of vtol transition
The next step in the sequence is transition to MC. By setting
vtol_back_transition we ensure that the acceptance radius is adapted to
the expected transition distance.
If 0, 1 and/or 2 file descriptors are not open when mavlink module
starts (as might be the case for USB auto-start), use default /dev/null
so that these numbers are not used by other other files.
Instead of interpret a request for "more logs than currently exists" as
a new request, use a request for index 0, which is more likely to be
the first request.
* [npfg]: Remove the guideToPoint function and replace with guideToPath
* [npfg]: remove unused navigateXXX functions
* [npfg]: Move navigateXXX Function into FWPoscontrol
* [FixedwingPositionControl]: Set default flaps and spoilers in attitude setpoint topic, and only change if necessary.
- precision landing works incorrectly, target position is not updated during the descent above target
- _prepareLandSetpoints needs to update _land_position continuously
Co-authored-by: kapacheuski <kapacheuski@gmail.com>
- the thermal offsets are an optional correction applied to the raw data, so when updating an existing calibration offset with new learned bias we don't want this incorporated
- EKF isTimedOut(), isRecent(), and isNewestSampleRecent() need to handle the case where the timestamp has never been set
- reset() more thoroughly reset fields (mainly impacts unit tests)
Set this flag to true if local position is valid but accuracy low, such that
the operator can be warned before system switches to position-failure failsafe.
Additionally, switch to RTL if currently in Mission or Loiter to try to reach home
or fly out of GNSS-denied area.
Set low accuracy threshold to 50m by default for FW and VTOL.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Otherwise the flight time restriction flag gets cleared too early, before
disarming (which puts the vehicle into the previous mode and it might
take off again).
Fixes the following case:
- user intention set to X
- failsafe triggers, mode = Y
- can_run for X becomes false
- user tries to switch to X
-> need to re-evaluate can_run
Previously, it was not possible to enable forwarding of messages to/from
teh USB instance because it does not have a param for it, like the
serial instances have.
With this commit, we change the default to always set forwarding on for
the USB instance as that is likely desired.
Signed-off-by: Julian Oes <julian@oes.ch>
- from the C standard, "All identifiers that begin with an underscore and either an uppercase letter or another underscore are always reserved for any use"
- from the C standard, "All identifiers that begin with an underscore and either an uppercase letter or another underscore are always reserved for any use"
The static object is destroyed on at_exit while threads might still be
inside a CS. This can lead to a hanging process.
Cleaner would be to gracefully stop the threads.
According to https://linux.die.net/man/3/pthread_cond_destroy:
Attempting to destroy a condition variable upon which other threads are
currently blocked results in undefined behavior.
Instead check if system has previously switched into LOITER to acheive the current
WP of type POSITION, and in that case stay in LOITER until altitude is reached.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Do the same as DO_REPOSITION wit only the altitude field populated
and MAV_DO_REPOSITION_FLAGS set, which means:
- switch to Loiter mode if not already in it
- set the current altitude to what is specified in the altitdue field,
keep current altitude setpoint otherwise
- keep current position setpoint
- fall back to current estimated position in case a position setpoint
is not finite
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Previously the condition was based on hard coded height rate estimate and
setpoint values and an altitude error threshold. That showed to be leading
to false positives when the vehicle doesn't tightly follow the altitdue
ramp given by TECS to achieve a new altitude setpoint, and has become
completely infeasibly with the latest TECS rework that leads to non-ramped
altitude setpoint changes in the tecs_status message.
The new check no longer checks the altitude error but only the height rate
instead. It begins to integrate the height rate error once it detects an
uncommanded descend condition (height rate negative while setpoint is
positive). Integral threshold can be tuned by user (VT_QC_HR_ERROR_I).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
In rare occasions asking for the protocol values after setting it returned
[0, 0]. I did not see any documentation for having to wait, but adding a
short wait period fixes it.
- refactor all EKF backend output predictor pieces into new OutputPredictor class
- output states are now calculated immediately with new high rate IMU rather than after EKF update
- IMU delayed sample is passed as around as control data to avoid storing an extra copy and make the requirement clear
- update motor mapping to use new UART_ESC_FUNC* auto generated params
- add support for actuator_test msg to support Actuator Testing in QGC
- modalai_fc-vX targets start driver if configured
- keep track of ESC spin direction in own param
- set ramp up param in MixerOutput to false
- use `gazebo-classic` everywhere consistently referring to the original Gazebo (eg version 9,10,11)
- additional `gazebo_*` helper targets added for compatibility, but warn about deprecation and tell you the new target naming
- use `gz` everywhere when referring to Gazebo (aka Ignition Gazebo or new Gazebo)
For the esc_calibration code, the actuator test is published for each motor
in a row. If the orb queue size is too small, messages are lost and not
received in mixer_module.
Set the default orb queue size of ActuatorTest high enough to keep the commands
for all motors in the queue.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Strictly follow the following convention for tailsitter:
FW Attitude and FW rate controller always operate in the FW frame, meaning that roll is
roll in FW, which for tailsitter means around the yaw axis in the body frame. The interfaces
between modules is though always in body frame.
That enables us to do the axis transformations for tailsitter, that are currently distributed
all over the controller (attitude, rate, vtol module), only at the input and output data of modules.
Side effect is that the FW rate control tuning gains meanings change: while before the roll gains
where meant for the body axis, they are now always applied for the FW roll axis (also in hover). So
the naming now is correct for FW, while before it was for Hover.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
This failsafe flag is currently used for not allowing to re-transition to FW, as well
as disabling pusher assist in hover. Till now it was only possible to reset it with
a commanded transition to MC, which many ground station interfaces didn't allow
as the system, after a quad-chute, already was in MC mode.
Hereby it is changed to reset when a new transition to FW is triggered (either via
RC switch or MAVLink command). It is the users responsibility to assess the situation
after a quad-chute happened to try to transition to FW again, manually proceed/land
the vehicle in MC, or let it finish the defined behavior after a quad-chute.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Note: This is the source of truth for the active maintainers of PX4 ecosystem.
See also [maintainers list](https://px4.io/community/maintainers/) (px4.io) and the [contributors list](https://github.com/PX4/PX4-Autopilot/graphs/contributors) (Github).
See also [maintainers list](https://px4.io/community/maintainers/) (px4.io) and the [contributors list](https://github.com/PX4/PX4-Autopilot/graphs/contributors) (Github). However it may be not up to date.
## Supported Hardware
This repository contains code supporting Pixhawk standard boards (best supported, best tested, recommended choice) and proprietary boards.
Pixhawk standard boards and proprietary boards are shown below (discontinued boards aren't listed).
For the most up to date information, please visit [PX4 user Guide > Autopilot Hardware](https://docs.px4.io/main/en/flight_controller/).
### Pixhawk Standard Boards
* FMUv6X and FMUv6U (STM32H7, 2021)
* Various vendors will provide FMUv6X and FMUv6U based designs Q3/2021
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.
Thomas Debrunner