We've come accross a case where a ROS node would consistently take something
over 800 ms until the first arming check request subscription callback was
triggered.
After the first sample, the callback always triggered within the expected
timeframe.
Therefore this patch allows for more time right after registration until
timing out.
* Update Zenoh-pico from 0.7.0 to 1.0.0
* Update the zenoh-pico version to use PX4/dev/1.0.0-px4
* Remove the rostopic and rt/ prefix
* Unlike zenoh-bridge-dds we were using, zenoh-bridge-ros2dds is now adding the rt/ prefix automagically.
Avoid flying into terrain using the distance sensor.
Enable through the parameter NAV_MIN_GND_DIST.
Only active during commanded descents with vz>0 (to prevent climb-aways),
excluding landing and VTOL transitions.
It changes the altitude setpoint in the triplet to maintain the current altitude
and republish the triplet. We also change the mission item altitude used for
acceptance calculations to prevent getting stuck in a loop.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
battery failures can occur upon arming when the load gets sgnificant.
In that case the safest thing to do is prevent a takeoff before anything
worse happens.
The wind variance can be reset to a value larger than the wind init
variance (e.g.: when the reset occurs when flying close to the N or E
axis). Constraining the variance after a covariance initialization would
artificially increase the correlation and could destabilize the filter.
* added support for resetting wind states to external observation
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* moved wind related functions into separate file
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* added VEHICLE_CMD_EXTERNAL_WIND_ESTIMATE
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* correctly compute variances
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* ekf2: implement wind reset
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* only allow VEHICLE_CMD_EXTERNAL_WIND_ESTIMATE during wind dead-reckoning
and increase horizontal velocity variance to allow velocity states to move
towards solution that is aligned with the newly set wind
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* only reset wind on ground
* still use wind reset using airspeed when it wasn't initialized
* exclude func for rover, change reset interface
* handle wind reset in drag-fusion
* replace state reset with variance reset in sideslip/drag fusion
* remove resetWind function
---------
Signed-off-by: RomanBapst <bapstroman@gmail.com>
Co-authored-by: Marco Hauswirth <marco.hauswirth@auterion.com>
- remove commander test ratio "tuning knobs" (COM_ARM_EKF_{HGT,POS,VEL,YAW})
- these are effectively redundant with the actual tuning (noise & gate)
in the estimator, plus most users have no idea why they'd be
adjusting these other than to silence an annoying preflight complaint
- remove ekf2 "PreFlightChecker" with hard coded innovation limits
- ekf2 preflight innovation flags are now simply if any active source
exceeds half the limit preflight
Also, as the flow makes the link between range and horizontal velocity,
only allow it to start if at least one of the two is known. Otherwise
the EKF will struggle to estimate both values at the same time.
* BMP581: Add Bosch BMP581 barometer
* Copyright：fix copyright header year
* style: not use pointers and Bool returns, Check the failed condition return
* delay: Replace usleep() with ScheduleDelayed()
* definitions: Delete unused definitions
* comment: Delet redundant comments
* constants: Change to uppercase
* BMP581: run make format
* estimator gps check fail flag for spoofing
* warn whenever spoofing state changes to true, use default hysteresis to completely stop fusion
* dont introduce more GPS namings, GNSS instead
* flash: exclude mantis for cuav_x7pro
- previously we could get stuck with optical flow still technically
active (_control_status.flags.opt_flow=true), but nothing being
updated due to excessive tilt, etc
- collapse updateOptFlow() and startFlowFusion() to avoid recomputing H
- this is a relatively expensive call we can easily avoid with the
right structure
- adjust flow sample gyro_rate immediately after popping from ring
buffer
- always update flow gyro bias (calcOptFlowBodyRateComp()) regardless
of flow quality or magnitude
* lib: add FilteredDerivative class
* AirspeedValidator: add first principle check
- filters throttle, pitch and airspeed rate, and triggers
if the airspeed rate is negative even though the vehicle
is pitching down and giving high throttle.
Check has to fail for duration defined by ASPD_FP_T_WINDOW
to trigger an airspeed failure.
* AirspeedValidator: define constants for first principle check
* FilteredDerivative: set initialised to false if sample interval is invalid
* airspeed_selector: improved comment
* increase IAS derivative filter time constant from 4 to 5
* use legacy parameter handling for FW_PSP_OFF
* handle FW_THR_MAX as well
* ROMFS/airframes: exclude some airframes for v6x and v4pro to save flash on them
---------
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Signed-off-by: RomanBapst <bapstroman@gmail.com>
Co-authored-by: Silvan Fuhrer <silvan@auterion.com>
* reset position by fusion
* handle local_pos_valid for fixed wing in gnss denied
* [WIP] ekf2: setEkfGlobalOrigin respect current height reference and vertical position aiding
* global origin, also reset vertical pos without gps as ref
* fix wo gnss, that bitcraze ci passes
* revert some changes as requested
* remove duplicate reset messages
* undo unrelated whitespace changes, I'll fix it everywhere in a followup
* [SQUASH] ekf2: add vehicle_command_ack
* resetGlobalPosToExternalObservation consolidate logic
* remove gnss check from local_pos validation check
* reset when 0<accuracy<1, otherwise fuse
* replace gps param with flag
* ekf2: dead reckon time exceeded, respect ZUPT preflight if air data or optical flow expected
* subtract timeout from last inertial dead-reck, change fake pos conditions, save flash
---------
Co-authored-by: Daniel Agar <daniel@agar.ca>
Pulls cell count, min voltage and max voltage from log file but still allows for over-rides. Also added debug info to tell user what what it found in the log and what it is using
Co-authored-by: chfriedrich98 <125505139+chfriedrich98@users.noreply.github.com>
Heading offset was configured as radians but should be configured as
degrees on Septentrio receivers. The parameter was already in degrees
but the configuration logic was changing it into radians. Also allow the
entire allowed range of heading offset values for Septentrio receivers.
The documentation for `SEP_DUMP_COMM` was quite unclear and users had to
use the user guide to find out what exactly it did. The new
documentation tries to make the purpose clearer.
This fixes an incorrect check of the device paths during instantiation
of the Septentrio driver that caused the driver to start and not print
an error message.
- track accel clipping count per axis
- only set bad_acc_clipping fault_status if at least one axis is
clipping continuously or if all have been clipping at warning level
- Note: this doesn't impact the clipping projections that boost the
accel process noise, pause bias estimation, and skip gravity fusion
on a per sample basis
- Fixes matplotlib version expression from ">=3.0.*" ro ">=3.0" which is the right syntax
Fixes issue #23329
Co-authored-by: lee wonwoo <leewonwoo@leeui-MacBookPro.local>
We had a case where someone took off with an experimental
system with 100% RAM usage on the embedded system
without noticing. This lead to problems during flight.
Since we already have a CPU load check it seems natural
to also check the reported RAM usage.
- new parameter EKF2_LOG_VERBOSE to enable (currently enabled by default)
- force advertise topics immediately (based on EKF2_LOG_VERBOSE and per aid source configuration)
- logger optionally log all estimator topics at minimal rate
When using optical flow for terrain aiding, we cannot assume that the
terrain estimate is valid if flow is fused as is can only be observed
during motion. When no direct observation is available, the terrain is
assumed to be valid if its 1sigma uncertainty is < 10% of the hagl.
Terrain is the state: terrain vertical position
Hagl (height above ground level) is the vertical distance between the
vertical position and the terrain vertical position
* mission: skip a vtol takoff mission item if already in air and a fixed wing
* MissionBase: also skip FW takeoff when already in-air
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
* mission: use setNextMissionItem to skip vtol takeoff when already in air
* mission: Only skip the VTOL takeoff in air for mission and rtl mission
If flying RTL mission reverse it must still include the takeoff point.
---------
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Co-authored-by: Silvan Fuhrer <silvan@auterion.com>
Without this flag the command silently succeeds even though the logs contains
an error. It's much more developer friendly to fail early in case of an error.
The log path is then also shown in the console output.
- MPC_VEL_LP: new velocity first order low pass filter (off by default)
- MPC_VEL_NF_FRQ/MPC_VEL_NF_BW: new velocity notch filter (off by default)
- MPC_VELD_LP: existing velocity derivative low pass filter, but I've dropped the remaining controllib usage
The case where the airframe maintainer wants to enforce a reset to airframe is
not covered anymore with the `param set-default` mechanism. For products based
on PX4 this is still required to ensure proper functionality after a major update.
Having a generic interface over the GPS drivers makes dedicated
functionality for each driver harder. Move the Septentrio driver into
its own module under the `gnss` driver directory, and let it have its
own parameters for only the functionality it requires. This also helps
with adding new features because they only need to be implemented for
the driver that wants it, simplifying testing.
- these are old RC switch configuration parameters that haven't been used for
years, but were hardcoded into old versions of QGC
- RC_MAP_RATT_SW, RC_MAP_POSCTL_SW, RC_MAP_ACRO_SW, RC_MAP_STAB_SW, RC_MAP_MAN_SW deleted
-
* ist8310: do reset before WHOAMI call
Apparently, the IST8310's WHOAMI register is writeable. Presumably, this
can get corrupted by bus noise. It is only reset if powered off for 30s.
Therefore, we do a reset before doing the WHOAMI.
* ist8310: improve comment
Co-authored-by: Beat Küng <beat-kueng@gmx.net>
---------
Co-authored-by: Beat Küng <beat-kueng@gmx.net>
Compared to GNSS, alternate position observation methods are less accurate
and thus generally not good enough to do airspeed validation with.
Airspeed validation is thus disabled if no GNSS fusion is happening.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
* AirspeedValidator: remove additional one second of hysteresis for triggering
innovation checks
- this check already uses an integrator and so adding more delay just makes
log analysis more difficult and does not really add any value
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* removed unnecessary conditions
Signed-off-by: RomanBapst <bapstroman@gmail.com>
* AirspeedValidator: only disable innov checks if ASPD_FS_INTEG is negative
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
* get rid of unnecessary check on innovation threshold parameter
Signed-off-by: RomanBapst <bapstroman@gmail.com>
---------
Signed-off-by: RomanBapst <bapstroman@gmail.com>
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Co-authored-by: Silvan Fuhrer <silvan@auterion.com>
This fixes various edge cases when input is set to both: RC and
MAVLink gimbal protocol v2.
Specifically:
- We no longer immediately reset primary control if there is no update,
otherwise this will reset immediately after an commands.
- Talking of commands: GIMBAL_MANAGER_CONFIGURE no longer switches
control to MAVLink but only an actual incoming setpoint command does.
- And incoming setpoint command only triggers UpdatedActiveOnce which
means we will check RC again afterwards and if there is big movement
switch back to RC. That's the intuitive thing to do until we have
better reporting about who/what is in control.
- Also, with RC we no longer always set us to be in control but only on
major movement.
192.168.0.x is often used by routers for WIFI/ethernet networks, and thus
can create conflicts.
This can happen for example if a companion is connected to the FMU via
ethernet and at the same time connects to a WIFI network as DHCP client.
VTOLs consume a lot more power in hover copared to fixed-wing fligt.
The remaining flight time thus should reset if one switches from MC to FW,
as otherwise it takes several minutes until the estimate goes down.
The argparse module has been builtin to Python since
Python 3.2, released in 2011 (see
https://docs.python.org/3/whatsnew/3.2.html). Further,
the argparse pip module has not been released or updated
since 2015, and lacks some of the features of the modern,
built-in argparse. Drop the pip installed version in
favor of the built-in version.
Signed-off-by: Chris Lalancette <clalancette@gmail.com>
- upate to NOAA grid API to build WMM table in one pass
- refactor declination/inclination/totalintensity table printing to
shared method
- compute scaling factor to maximize resolution
fix: ControlAllocator float and int comparison bug
There was an incorrect comparison between a float variable `dt` and a `uint64_t` value representing 5 milliseconds (`5_ms`). As a result, `do_update` could never become true even if the last torque setpoint was received more than 5 milliseconds before.
To solve this, the `5_ms` value has been converted to seconds (0.005f) for the comparison with `dt`.
In this mode, the mag is used to inilialize the heading.
During flight, the heading is predicted using gyros and corrected
by GNSS measurements if available.
Previously, if a mag was not required (not index 0 and not used by ekf),
it was not counted in num_enabled_and_valid_calibration.
If a user set SYS_HAS_MAG to e.g. 3, it would then trigger a preflight
failure, even if there were 3 calibrated and enabled mags.
The EKF can recover from an initial bad earth mag field estimate.
Constraining the field is not necessary and can lead to an unpredicted
behavior of the filter.
Declination fusion is safe to run even when the horizontal field is 0
On slowly moving vehicles (e.g.: boats, rovers), the yaw estimator has
worse convergence than the main EKF. Resetting the mag states using the
yaw estimator as reference can lead to poor heading. Also, the EKF can
recover really well from initially incorrect mag states.
The previous default of 900m leads to many warnings if left
unchanged, especially if the vehicle is already in-air when
the Mission is started.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Includes:
- Remove some of the outdated Python2 checks and compatibility.
- Try not catch all exceptions but only the expected ones. Otherwise,
this makes it really hard to debug if anything unexpected actually
goes wrong.
- Make use of fstrings.
- Make output slightly prettier.
Signed-off-by: Julian Oes <julian@oes.ch>
This adds a new protocol extension which allows to get the bootloader
version.
The bootloader version is different from the bootloader protocol
revision which has stabilized at 5 and is not easy to update unless a
bootloader is actually breaking the protocol. The reason being that both
the Python script as well as the uploader used in QGC will not attempt
to load firmware if they don't know the bootloader version, so it could
basically be considered a "breaking" protocol revision.
Signed-off-by: Julian Oes <julian@oes.ch>
If the STM32H7 fails to program or erase a full chunk of 256 bytes, the
ECC check will trigger a busfault when trying to read from it.
To speed up erasing and optimize wear, we read before erasing to check
if it actually needs erasing. That's when a busfault happens and the
erase time outs.
The workaround is to add an option to do a full erase without check.
Credit goes to:
https://github.com/ArduPilot/ardupilot/pull/22090
And the protocol option added to the bootloader is the same as for
ArduPilot, so compatible.
Signed-off-by: Julian Oes <julian@oes.ch>
* usb: Added parameter to enable always starting mavlink on USB.
Refactored cdcacm_init into a module and added a paramter to allow always starting mavlink on
USB, also added a paramter to choose the mode. The current default behavior is to wait and listen
for data on USB and auto-detect the protocol (mavlink, nsh, ublox). This results in the mavlink
stream not starting until something else on the mavlink network sends a packet first. The new
default behavior is to always start mavlink.
Added parameters
MAV_USB_ENABLE -- default 1 (always start mavlink on USB)
MAV_USE_MODE -- default 3 (onboard)
* added 3 retries for opening serial port in mavlink, removed sleep before sercon
* added DRIVERS_CDCACM_AUTOSTART to ark-v6x default.px4board
* added CONFIG_DRIVERS_CDCACM_AUTOSTART=y to default.px4board for boards with CONFIG_CDCACM in their nsh/defconfig
* format
* remove PGA460 from COMMON_DISTANCE_SENSOR to save flash
* remove LIS2MDL from COMMON_MAGNETOMETER to save flash
* disable CONFIG_DRIVERS_CDCACM_AUTOSTART for fmu-v5 protected.px4board
* moved and renamed parameters, removed mode logic in mavlink
* changed parameter names, added mode none
* remove parameters from mavlink
- increase symforce CppConfig zero_initialization_sparsity_threshold so
that a Matrix setZero() call is performed instead of individually zeroing
Co-authored-by: bresch <brescianimathieu@gmail.com>
The setting wasn't consistent with the one used in the Visual Studio
Code settings, which caused different newline formatting depending on
whether the user uses Visual Studio Code or another editor that uses
EditorConfig.
This notes the reference yaw angle for the Septentrio Mosaic-H.
It's unfortunately a bit tricky in that Unicore has the main antenna
in front by default while Septentrio decided to put the aux antenna in
front.
Signed-off-by: Julian Oes <julian@oes.ch>
Regression from https://github.com/PX4/PX4-Autopilot/pull/23043
Also avoids a race condition by making sure the command ack is handled
before sending out the mavlink message (in case an external component
reacts immediately to the mavlink message).
- this warning was to catch any potential errors in sensor selection
relative to what's actually available, we don't need to complain
about initial selection before the EKF selector is available
2024-05-02 11:53:31 -04:00
998 changed files with 28688 additions and 12601 deletions
# Check if there are files checked in that don't end in a newline (POSIX requirement)
git grep --cached -Il ''| xargs -L1 bash -c 'if test "$(tail -c 1 "$0")"; then echo "No new line at end of $0"; exit 1; fi'
if[$? -ne 0];then
return_value=1
fi
# Check if there are files checked in that have duplicate newlines at the end (fail trailing whitespace checks)
git grep --cached -Il ''| xargs -L1 bash -c 'if tail -c 2 "$0" | ( read x && read y && [ x"$x" = x ] && [ x"$y" = x ]); then echo "Multiple newlines at the end of $0"; exit 1; fi'
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.
Ramon Roche