Attendee count (max): 18
- Copter and Plane 4.2 are out!
- Tracker 4.2 soon
- Sub… l… 4.1 first?
- Blimp too….
UTC1101 - Rover/Copter update
- Releases went out
- Caveats on 4.2 release
- Combinations of ESCs and boards which aren’t working with d-shot
- Were working with 4.1.5
- Andy’s working on it
- Serial1 vs serial2 is a bit of an issue
- GPIO/change to servo function - -1 thing
- Precision landing retry 3 times
- 4.2.1 not too far away
- Some patches in Plane 4.2.1 which are candidates
- Probably do the double-rebase thing
UTC1106 - Plane update
- 4.2.1 is out
- Important land-reposition issue fixed
- Race condition disarm
- extremely unlikely to trigger, been in since forever
- Has to include a double-landing to happen
- Need to check Copter
- Maybe with MIS_RESTART false
- Motor-offset parameter for reducing CAN bandwidth
- 32-servo support merged
- Flown on 3 vehicles
- Multirotor (inc. dshot)
- Should we look
UTC1113 - Should we look at the roadmap each month?
- We sometimes forget what we’re heading towards
- Want major 4.3 features early in release cycle
- E.g. update ChibiOS sooner rather than later
UTC1115 - Add option to select branches by MichelleRos · Pull Request #32 · ArduPilot/CustomBuild · GitHub
- Ability to test custom build server
- 4.2 we dropped some important features for some boards
- E.g. optical flow on F4
- Need to be able to compile for the stable release
- Really looking forward to AP_Periph being able to be compiled on here
- Could run a custom-beta.ardupilot.org to allow testing of different branches
- Label on github, use github API to get things with the relevant tag?
- Adds to dropdown list
- Might need to whitelist
- [9:27 AM] Peter Hall: I suspect we could use a CI to remove a label each new push
- Need to be smart about restricting which vehicles can be built from which branches?
UTC1120 - The document of nokov module by 15871812598 · Pull Request #4356 · ArduPilot/ardupilot_wiki · GitHub
- Add support for nokov motion capture support
- There’s a MAVProxy PR to go along with this Wiki document
- Missing some files in the MAVProxy
- Worried that nobody else has tested it…
UTC1144 - AP_Baro: tidy probing of external i2c baros by peterbarker · Pull Request #19917 · ArduPilot/ardupilot · GitHub
- Makes probing of external barometers more uniform
- Henry’s never managed to get an external baro working on SpeedyBeef
- Josh will put this on hardware to test it
UTC1154 - AP_UAVCAN: add option to send ESC msg idx to match motor name idx by magicrub · Pull Request #19967 · ArduPilot/ardupilot · GitHub
- Add option to send ESC msg idx to match motor name
- Uavcan motor2 can be index 2
- This PR would allow for ambiguity with things like throttle out
- How does this interact with OF_OFFSET?
- OF_OFFSET helps really reduce bus bandwidth
- Can-dshot periphs work without changing
- 5,6,7,8 with offset 3 would save 25%
- Tridge is thinking of a new message with 11 bits of throttle value
- The usual 2000-ish throttle steps
- 5 bits for an index number….
- This approach expects everything is coming from AP_Motors and its motormask/numbering
- ESC telem on quadplane with motors on 5,6,7,8 would report as rc channels 1,2,3,4
- This was really confusing
- ESC feedback isn’t intuitive
- They map to the servo number now
- User has to know the servo number of each actuator (e.g. SERVO3_FUNCTION)
- ESC telem 3 is then that function
- No mapping directly to motors, ‘though
- We really should have called them actuators
- OUT_ on AP_Periph
- Much better name for other vehicles
- Big rename would cause lots of issues
- Helicopter outputs are servos and ESCs…
- The solid 1-1 esc-telem to servo output number is concrete and easily to explain
- This would be baffling for any moderately weird vehicle
- We should add the mavlink messages for the higher-numbered actuators
- ESC_TELEMETRY_13_TO_16 etc up to 32
- Motor 1 2 3 4 is something the user can understand
- Inherent conflicts with this approach….
- [9:59 AM] Peter Hall: now we have 32 servos we have more flexibility to have motors in a sensible place and order.
- [10:01 AM] Peter Hall: AP motors does not always use motor outputs either.
- [10:02 AM] Peter Hall: Copter bicopter uses throttle left and right.
- [10:03 AM] Peter Hall: Tricopter uses motor 7 as a servo output not a motor.
- [10:03 AM] Peter Hall: There are a few odd cases that brake the servo motor function to motor number rule.
- [10:09 AM] Peter Hall: Two layer hex is a bad example, its a test order to motor number mapping IRRC
- [10:09 AM] Peter Hall: We have been doing that for all the recent ones.
- [10:11 AM] Peter Barker: @Peter Hall I picture a pretty 3D picture of a 2-layer (12 motor!) hexacopter, and the GCS being able to highlight the bad motor in red.
- [10:11 AM] Peter Barker: Would be nice to be able to provide enough information for the GCS to be able to do that…
- [10:12 AM] Peter Hall: Right, 2 layer hex is starting front right, top → bottom clockwise.
- [10:12 AM] Peter Hall: The motor number is the testing order.
- [10:12 AM] Peter Barker: … same thing with motor test - being able to highlight the motor that’s supposed to be spinning on a little picture would be awesome, right?
UTC0015 - Avoidance: Path planning path reporting by IamPete1 · Pull Request #20554 · ArduPilot/ardupilot · GitHub
UTC0028 - Update OSD flightmodes for plane and copter by andyp1per · Pull Request #20811 · ArduPilot/ardupilot · GitHub
- Add extra flightmodes into OSD
- Avoids numbers in interface
- Any PR should mention what testing has been done on it!
- Fix at same time
UTC0029 - Add W25Q128 variant by andyp1per · Pull Request #20817 · ArduPilot/ardupilot · GitHub
- Add W25Q128 variant
UTC0032 - Add support for Plane in SITL-on-HW by peterbarker · Pull Request #20822 · ArduPilot/ardupilot · GitHub
- Sim on hardware for Plane
- Also allows mask for actuator output
UTC0038 - Allow SITL rangefinder to work on sim-on-hardware by peterbarker · Pull Request #20825 · ArduPilot/ardupilot · GitHub
- Simulated rangefinder support for sim-on-hardware
UTC0039 - GSoC update
- Announcements have been made
- Shiv, Randy and Peter have arranged a discussion in Discord
- PH is going to work with Simulink
- Sanket with Jaime and revisiting ROS+ArduPilot integration
- James’ comments very valid
- Huge issue that we can’t integrate with ROS2 ATM
- MAVROS doesn’t work with it
- Maybe there’s a branch somewhere?
- All news to Randy
- Randy’s understanding
- Can’t use MAVROS with ROS2
- Direct integration was done last year?
- Project fell apart somehow
- Massive change in ArduPilot
- Lack of developer time
- ROS2 protocol
- Up-and-running on SITL
- Another major protocol besides mavlink
- Huge, complex protocol
- Wasn’t up to a mergeable standard
- Basic operations in SITL worked
- Not in good-enough shape to merge
- Current plan is to go with ROS1
- March along this path?
- Perhaps see if MAVROS has been done for ROS2…
- ROS2 is a completely different sort of protocol to ROS1
- [10:43 AM] Buzz: Porting MAVROS for ROS2 (GSOC'21) <<-discussion here…
- ArduPilot Discourse
- Porting MAVROS for ROS2 (GSOC’21)
- Ardupilot recently updated their GSOC idealist and I was interested in working on the project Mavros for ROS2. I have some queries related to the project and it would be great if some of the possible mentors could answer them : 1.) As of now I have a basic PixHawk kit ready and set up and do I need some additional hardware like Lidar Sensors as…
- [10:44 AM] Buzz: “There is almost ready port of mavros to ROS2 (GitHub - uas-at-ucla-dependencies/mavros 56), it need some cleaning and testing. AFAIR I manage to run it.”
- Ideally we do ROS2 as James says
- Could we clean up the ROS2 direct interface that was done last year?
- Would seem to be a better path than MAVROS…
- Dds protocol
- Might be possible to do this in LUA….
- Links against AP_XRCE submodule library
- Thinks it is running on POSIX
- Clock_gettime etc etc
- Current direct implementation really needed some sort of generator/compiler as it was doing a lot of assembly by hand
UTC0101 - close