* fix(zenoh): generate the topic catalog into the build tree
Kconfig.topics is fully generated from the uORB message set, but the build
regenerated it into the source tree at configure time. Building any Zenoh
board therefore left the working tree dirty whenever the message set had
changed since the file was last committed (the catalog is board-config
dependent, so it drifts easily).
Generate the catalog into the build directory in cmake/kconfig.cmake, before
Kconfig is parsed, and source it from there via ZENOH_KCONFIG_TOPICS. It is
generated board-independently from every message so it no longer depends on
the msg-gating Kconfig symbols it is sourced alongside; the per-board factory
still gates which topics are actually compiled. Drop the committed catalog.
* chore(zenoh): bump zenoh-pico to the build-tree header fix
Moves zenoh-pico's generated config.h/zenoh-pico.h/library.json out of its
own source tree and into the build tree, so building no longer dirties the
submodule.
Gated on PX4/zenoh-pico#2: the pointer currently references the fix branch on
a fork and must be moved to the merged commit before this is ready.
* bump zenoh-pico
* feat(failure_injection): integrate failure injection support across sensor drivers
* feat(failure_injection): enhance failure injection with RC switch support and instance bitmasking
feat(failure_injection): add disabled failure injection manager and system command support for v5x and v6x boards
* feat(failure_injection): add battery failure injection
Add a value-mutating apply-site for FAILURE_UNIT_SYSTEM_BATTERY. On an
injected OFF the outgoing battery_status is reported as a depleted pack
(zero remaining, emergency warning) so the low-battery failsafe triggers.
The apply-site lives in the shared Battery library, covering the analog
ADC, INA power monitors, ESC battery and SITL in one place, plus the
UAVCAN battery driver which publishes battery_status directly. The
previous SITL-only hack in BatterySimulator is removed in favour of this
shared path so simulation and hardware behave identically.
* fix(failure_injection): change parameter types from int32 to enum
* refactor(failure_injection): disable failure injection manager and system commands across multiple boards
* feat(failure_injection): enhance failure injection with timestamp handling and message-less support
* refactor(failure_injection): simplify has_timestamp_sample implementation and remove unused includes
* refactor(failure_injection):move conditional compilation into helper libary
* refactor(failure_injection): update CMakeLists to include failure_injection dependency across multiple drivers
---------
Co-authored-by: Claudio Chies <chiesc@chies.com>
The cdrstream uORB->IDL->CDR codegen floods the build log with noise:
- CycloneDDS idlc warns once per carried-over .msg comment that the
@verbatim annotation is unsupported (VehicleCommand alone emits 153).
Pass -Wno-unsupported-annotations via the idlc_generate WARNINGS list.
- rosidl_adapter prints a Reading/Writing line per message to stdout.
Filter those in msg2idl.py via builtins.print rather than redirecting
stdout, which the empy template engine breaks on. Errors still go to
stderr.
* fix(boards/ark/fpv): split FW/VTOL into separate builds for flash
Default ark_fpv is multicopter-only so listener and sd_bench fit in
flash. Fixed-wing and VTOL stacks move to ark_fpv_fw and ark_fpv_vtol,
which disable those two system commands to stay under the limit.
Signed-off-by: alexklimaj <alex@arkelectron.com>
* fix(boards/ark/fpv): trim optional modules from vtol build
Drop gimbal, gyro_fft, autotune, camera, payload, and debug systemcmds
from ark_fpv_vtol so the full VTOL stack fits with more flash margin.
Signed-off-by: alexklimaj <alex@arkelectron.com>
* fix(boards/ark/fpv): keep more modules on vtol; trim only debug extras
Restore autotune, gimbal, gyro_fft, mag bias, payload, camera, esc
battery, and actuator_test on ark_fpv_vtol. Keep flash margin by
leaving sd_bench, listener, bsondump, i2cdetect, and pca9685 off.
Signed-off-by: alexklimaj <alex@arkelectron.com>
---------
Signed-off-by: alexklimaj <alex@arkelectron.com>
* chore(claude): overhaul review-pr skill to focus on substance
The skill spent most of its 207 lines on process: merge-strategy
recommendations, commit-message auditing, formatting checks, and an
interactive posting menu. That feedback is nit-picky noise for a PR
author, and CI already enforces formatting.
Rewrite it (49 lines) around reviewing substance: merit and need (root
cause vs papering over), first-principles analysis for physics/math
changes, architecture consistency with the surrounding code, and an
alternatives assessment. Output is a debrief for the operator plus a
terse, actionable draft comment explicitly attributed as a Claude
review.
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
* chore(claude): generalize numeric-robustness check beyond float32
The physics/math check hardcoded float32; not all flight-critical math
uses it. Phrase the robustness check in terms of the given
representation, keeping float32 precision as a concrete example.
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
* chore(claude): fold maintainability into the architecture check
Rename the check to Architecture and maintainability and direct it at
hidden cross-module coupling — a change that silently breaks an
assumption in another module — asking for a contract-codifying unit
test when such coupling is unavoidable.
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
* chore(claude): add reliability to the correctness check
Reliability was not called out anywhere. Rename to Correctness and
reliability and add resource exhaustion and failure-path handling to
the list of defects to look for.
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
* chore(claude): make the debrief lead with findings
A per-file restatement of a well-described PR wastes tokens and time.
Have the debrief lead with actionable findings and only explain what
the change does when the PR description is missing, ambiguous, or
misleading.
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
* chore(claude): require provable, objectively-framed findings
Direct the draft comment to frame findings as engineering tradeoffs
rather than judgments, and to stay silent on any flaw it cannot
demonstrate via a concrete code path or first-principles derivation.
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
---------
Signed-off-by: Jacob Dahl <dahl.jakejacob@gmail.com>
Strict command-parameter validation added in #27541 rejects
MAV_CMD_NAV_TAKEOFF when an unsupported param is set. QGroundControl
sends param1 (minimum pitch) = -1 on multicopter takeoff, which is
non-zero/non-NaN, so commands were silently denied (MAV_RESULT_DENIED
with no log) on MC airframes and QGC guided takeoff stopped working.
Extend the NAV_TAKEOFF override to include VEHICLE_MC so param1 is
accepted and ignored on multicopters, matching FW/VTOL handling.
Homebrew 6.0+ refuses to load formulae from untrusted third-party taps. This adds a trusted block for OSRF taps.
Signed-off-by: Cade Andersen <cadecandersen@gmail.com>
In AUX output mode, the gimbal module unconditionally acted as a
virtual gimbal device and therefore advertised a gimbal manager, even
if no output channel was assigned a gimbal output function, i.e. no
gimbal could possibly be connected. This conflicts with setups where a
gimbal manager external to PX4 is used.
Only act as a gimbal device (and hence advertise a gimbal manager) if
at least one output function param is set to gimbal roll/pitch/yaw,
mirroring the MAVLink v2 output mode which only advertises once a
gimbal device is discovered. The check is done once at module startup
to avoid scanning all params on every parameter update, so assigning
a gimbal output function requires a reboot to take effect.
Signed-off-by: Julian Oes <julian@oes.ch>
* feat(navigator): extend detect and avoid module to follow regulatory standards such as ASTM F3442
* docs(docs): minor subedit
* refactor(navigator): reduce flash by grouping notif into same events
* docs(daa): Improve docs readability with ::: details blocks
* fix(navigator): single event when on ground with conflict
* refactor(boards): increase flash length from 4M - 128k to 5M - 128k
* rework(general): define DAA standard at build with new CONFIG_NAVIGATOR_ADSB_F3442
* clean(navigator): minor changes to clean the PR
* feat(boards): add CONFIG_NAVIGATOR_ADSB_FAKE_TRAFFIC in visionTargetEstStatic.px4board
* rework(daa): reduce amount of abstractions and minor cleaning
* refactor(boards): allyes revert flash length to 4M-128K
* refactor(boards): allyes increase flash length to 5M-128K
* rework(daa): rework event notifications to improve clarity
* docs(docs): move details inside of ::: details block
* docs(docs): run npx prettier
* refactor(daa): avoid void mutator functions
* docs(docs): Improve Python helper to decode daa unique id
* rework(daa): move encoded id handling to the adsb lib and refactor on_active
* refactor(daa): naming and zero init
* fix(daa): move dataman dep from daa level to unit test level
* refactor(daa): define common daa_input and rework process_transponder_report
* fix(daa): remove stale todo
* refactor(daa): rename crosstrack_based_daa to crosstrack_standard
* refactor(daa): rename F34_ params to DAA_
* docs(docs): revert changes to autogenerated docs
* docs(docs): Add Detect And Avoid in index.md
* refactor(daa): update message on init failed
* refactor(daa): only publish most_urgent conflict once in on_active()
* refactor(daa): move conflict buffer handling in the adsb lib
* refactor(daa): move notifications into new class ConflictNotifier and only notify once per cycle
* refactor(daa): uninit _cycle_changes to store into .bss (and save flash)
* docs(docs): remove unused image link
* refactor(daa): move automated action policy to the adsb lib
* refactor(daa): move self detection into adsb lib as DaaTrafficFilter
* refactor(daa): minor changes in unit tests
* refactor(daa): merge DaaTrafficFilter and DaaEncoding
* fix(daa): fix CI by removing navigator from the DAA deps
* refactor(daa): reduce stack size
* fix(daa): advertise _fake_traffic_pub in constructor
* refactor(daa): Comments only, simplify and reduce comments
* fix(daa): brief comment missing end star
* Disable CONFIG_NAVIGATOR_ADSB in ark_fpv_default board
* docs(docs): General clarifications and remove the 1.18 badge
* refactor(daa): cleaning
* refactor(daa): minor cleaning
* refactor(adsb): simplify conflict tracker assuming push_back cannot fail
* build(cmake): cleaner fix to protect regex alternations
* refactor(daa): Crosstrack, remove unit test requiring finite yaw (yaw not used by standard)
* refactor(daa): minor cleaning and fix callsign to uint64
---------
Co-authored-by: jonas <jonas.perolini@rigi.tech>
Co-authored-by: Hamish Willee <hamishwillee@gmail.com>
* docs(mc_pos_control): fix MPC_ACC_HOR_MAX description for MPC_POS_MODE
MPC_POS_MODE only exposes Direct velocity (0) and Acceleration based (4).
Replace the stale "mode 1" wording so parameter metadata matches the live enum.
* docs(mc_pos_control): shorten MPC_ACC_HOR_MAX long description
Keep it short per review (@dakejahl), in the spirit of #27758.
Update the mavlink receiving example to match the current behavior from
"b52464059e fix(mavlink): preserve invalid battery values on receive"
patch.
Signed-off-by: Onur Özkan <work@onurozkan.dev>
* feat(driver/heater): Add threshold to prevent the heater from always heating.
It only starts heating if the temperature drops below the specified threshold,
and continues until reset.
Battery::updateDt() clamps _dt to a maximum of 2 seconds:
_dt = math::min((timestamp - _last_timestamp) / 1e6f, 2.f);
This same _dt is then used for two different purposes:
1. Battery::sumDischarged(), the coulomb count that accumulates
discharged_mah: discharged_mah_loop = current_a * dt.
2. The current average low-pass filter update
(_current_average_filter_a.update(current_a, _dt)).
The 2-second clamp makes sense for (2): a low-pass filter update
generally assumes a roughly steady sample interval and shouldn't see a
single very large dt distort its state. It does not make sense for
(1): the coulomb count is a plain physical integral (charge = current
x time), and clamping dt there just means an update that arrives less
often than every 2 seconds (e.g. some DroneCAN/UAVCAN battery devices,
as reported in #27797) has its actual elapsed time silently truncated
to 2 seconds, systematically underestimating discharged_mah for that
update - the error accumulates over the life of the battery.
Fix: track a second, unclamped dt (_dt_discharge) alongside the
existing clamped one, and use it specifically for the coulomb count in
sumDischarged(). The filter update keeps using the existing clamped
_dt unchanged.
Fixes#27797
Test plan:
- I don't have hardware with a >2-second-interval battery update
source to reproduce this end-to-end, so I verified the corrected
arithmetic with a standalone C reproduction of updateDt()/
sumDischarged(): a 5-second update interval at a constant 10A
before the fix accumulates 5.56 mAh (current x 2s, clamped) instead
of the physically correct 13.89 mAh (current x 5s); after the fix it
correctly accumulates 13.89 mAh.
- Verified a normal high-frequency update pattern (10 x 100ms updates
at 5A) is unaffected: both before and after the fix accumulate the
same, correct 1.39 mAh for that 1-second span.
Signed-off-by: yi chen <94xhn1@gmail.com>
Co-authored-by: yi chen <94xhn1@gmail.com>
In navigator_main we use needsStraightLineFallback do determine whether
or not to show the error "RTL: no geofence avoidance path; flying directly".
Without any geofences that is not an error but expected behaviour.
Fix: When there are no geofences, do not declare
needsStraightLineFallback. Behaviour is otherwise the same - the geofence
avoidance path is empty, hasMore() always false, and we return directly.
This is the only double-precision exp in the codebase, and occupies a
hefty chunk of flash.
Relevant history:
- https://github.com/PX4/PX4-Autopilot/pull/9365 / 39bb65ffd7
- Timesync algorithm is initially added with this "sigmoid"
function calculated completely in float
- but _filter_alpha / _filter_beta are double, presumably to be
compatible with adjusting _time_offset in add_sample which is a
double (which it needs to be to represent large offsets coming
from int64_t)
- https://github.com/PX4/PX4-Autopilot/pull/9809 / cf74166801
- to silence warnings, the expf is swapped for an exp and the
entire sigmoid calculation is now in double
I would say the better fix to silence the promotion warning is to keep
the sigmoid function in float (precision not needed) and cast to double
right after, to not accidentally change anything depending on double.
Replaces the cube-root pow() with cbrt() (still double precision, no
log lookup table) and the pow(x, 2) calls with x*x. Removing the last
double pow() callers on the target lets the linker drop __pow_log_data
and __exp_data (~6.3 KB flash) with no loss of geodetic precision.
The ~PublicationBase() destructor was header-inline, so its null-check
plus the orb_get_queue_size()/unadvertise() branch was emitted at every
publication destruction site (PX4 links with bfd ld, no ICF, no LTO).
Move it into Publication.cpp alongside the already out-lined
advertise()/publish(), mirroring #27581.
Saves 2.4 kB of flash on px4_fmu-v6x_default.
publish(), advertise() and get_instance() were instantiated per topic
type and the advertise-check in publish() was additionally inlined at
every call site. Move them into a type-independent PublicationMultiBase
compiled once.
Saves 5.5 kB of flash on px4_fmu-v6x_default.
* refactor(simulation): use single-precision noise in SensorBaroSim
The Box-Muller noise generator ran in double precision, pulling in the
double log() implementation and its __log_data table (~2.2 KB). Noise
generation does not need double precision; converting the block to
float drops the last double log() caller on the target.
* fix(simulation): remove double Box-Muller in SensorBaroSim noise
generate_wgn() already returns a standard normal sample (Marsaglia polar
method), but the pressure-noise block fed it into a _second_ polar
Box-Muller transform as if it were a uniform [0,1) source.
Fix: Use the generator output directly for the intended 1 Pa RMS N(0,1)
noise and remove the second noise transform.
Plan RTL paths that route around geofence boundaries — both inclusion
and exclusion zones — instead of flying straight through them and
breaching. The planner builds a visibility graph over the margin-inflated
geofence polygons and circles and runs Dijkstra to find the shortest
legal return path, falling back to a straight line to the destination
when no valid path exists.
Highlights:
- New reusable libraries: src/lib/dijkstra (generic shortest path) and
src/lib/geofence (fixed-point geometry, polygon inflation, bitangent
visibility); the RTL planner lives in navigator/RTLPlanner.
- Visibility graph keeps only bitangent edges and skips edges that poke
into a forbidden region, greatly reducing edge-cost computation.
- Corner splitting is limited to sharp convex corners.
- Geometry validation: reject self-intersecting polygons and vertices
outside the fixed-point range; centimeter fixed-point scaling keeps
orientation tests exact and avoids drift at large distances.
- Failure handling: a Status enum replaces silent bool returns, and
failures (unbuildable fence, planner capacity overflow, dataman load
errors, NaN waypoints, destinations that breach the fence) are
surfaced to the operator via MAVLink warnings/criticals; the planner
falls back to a straight-line RTL.
- Destination updates are centralized in RTL::setRtlTypeAndDestination;
the path is replanned only when the destination or geometry changes.
- VTOLs always use the fixed-wing margin (FW loiter radius).
- kMaxNodes is exposed as a Kconfig option (default 100); the feature is
disabled on flash-constrained boards (FMUv4 and older, various F4/F7).
- Tests: unit tests for the Dijkstra lib, geofence geometry utils, and
the avoidance planner, plus a MAVSDK SITL test flying an RTL through a
geofence.
- Documentation: new "Geofence Awareness" section in the RTL docs.
Co-authored-by: Balduin <balduin@auterion.com>
* docs(releases): fix malformed SIH wind PR link in v1.18 notes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): draft v1.18 release notes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): mark v1.18 beta and add changes merged since the draft
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): add v1.18 upgrade guide, hardware fixes, and section structure
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): resort v1.18 major changes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): link sensor and peripheral docs pages in v1.18 notes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): describe new boards and highlight serial passthrough in v1.18 notes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): fold standalone sensors section into hardware and common
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): fill v1.18 ethernet section with networking fixes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): merge ROS 2 and middleware sections with DDS and Zenoh subsections
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): group vehicle-type notes under vehicle-specific section
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): reorder v1.18 other-changes sections by topic group
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): drop other-changes wrapper heading in v1.18 notes
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): organize debug-logging subsections and add infrastructure section
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* docs(releases): fix main.md template wording and mark v1.18 beta in sidebar
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
---------
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
