PX4 Dev Call: September 16, 2020

September 16, 2020

Agenda

  • Status update by component/project
  • Roadmap, and Release discussion
  • Community Q&A
  • In-Depth discussions

Join Meeting

Meeting ID: 946 175 205
Join using your mobile/desktop

Call in using your phone, find your local number: https://zoom.us/u/aetSYKbiMF


Component update

UAVCAN / Hardware Workgroups

Pavel was on the call:

  • Move DS-015 documentation from Google docs to Latex on GitHub
  • Physical layer connectors were discussed
  • Pavel is going to create the message definitions for the ESCs
  • Discussion about options for the bootloader

System Architecture

OS / NuttX

blocked by:

  • Cygwin CI build, don’t block @MaEtUgR will create a follow-up pr to fix it
  • Makefile build is failing

Driver

Heater driver for px4io GPIO pin controlled heater available on cubes.

Commander

Failure detector improvements merged:

Failure detector extension draft to detect rate control divergence before flipping the vehicle:
It would be nice to have broken models in simulation to test this in CI.

There are flags from the power module about over current. We should plan acting on those.
@dagar has power module logic separated from ADC and based on that we could make sure the flags are correctly handled:

Estimation

EKF2 selector is being worked on by @Paul_Riseborough and in his latest test he had it working and switching over correctly when there a high gyroscope offset. With two estimators (and two IMUs) it takes ~0.5-1 second to switch based on the variance checks. With three estimators (and IMUs) it switches almost immediately because the comparison is easier. The pr is getting close to being mergeable. We are missing more testing and some flash. Also the default case of only having one EKF needs to be verified to work as expected with all sensor cases.

Question by @david_s5 : In the case that sensors change e.g. I2C address how can we detect that?

VTOL

No updates

Fixed Wing

@sfuhrer Bug report for TECS that we have a bugfix for. The fix was reviewed and test flown today:

Pitch offset confusion needs to get gets settled in TECS. There was discussion recently to close this out.

Hardcoded maximum of 4m/s airspeed reading for arming to pass the preflight checks. This gets parameterized to allow arming in high wind. There’s a discussion to improve the preflight check without adding more options that people will likely not know about. There should be the option to disable any preflight check in one single simple way and the value to make airspeed fail could be derived:

Multicopter

We need to test if it works as expected. E.g. with Yosemite gazebo world. The cases

Avoidance

No updates

Simulation

New simulator integration with JSBsim. It’s different from e.g. Flight Gear because it’s directly integrated into the simulator and not dependent on rendering speed. JSBsim has good flight dynamics but also requires a numerical model. It misses options for robotic simulation e.g. cameras. Discussion about what simulator(s) should be default and supported versus spreading the effort.

Replace xacro with jinja templates:

MAVSDK

MAVROS / DDS / ROS2

No updates


Release

Community Q&A

In-Depth discussions

  • Ethernet configuration
    Static IP address set on built time? Shouldn’t it be configurable? Depends on the use case. Product with fixed companion computer vs development setup.

Errata and Feedback

Let me know below if we failed to capture anything the right way, and if there are any updates (not present here), or you have feedback you would like to share with the dev-team.