- Community Announcement
- Community Q&A
- General Discussions
- Specify what you are trying to achieve
- Specify what environment / platform you are using
- Prepare a Link or Document to help understand if possible
- First, ask the question on Discord or create a Github Issue!
- If you take over 5 minutes for the question, please continue in Discord or a Github Issue.
ICM sensor data are clipping out a lot, could it be filtered somehow?
- Currently clipping at around 16G
- BMI sensor isn’t clipping.
- Some settings for the sensors have changed, may be realted to the result
- You can reduce the priority of the IMU that’s clipping to make it secondary, not primary, then:
- We could make it to have only the IMU that is actually being used to complain about clipping.
- Could be even better to place Flight Controller on the outskirt of the frame to reduce vibration on it.
- There also could be hardware hitting the Flight Controller, causing high acceleration
- Check out clip count that is logged, and if it is a sustained clipping, and not sporatic, it could cause flyaway
ICM42649 chip does it exist?
- Probably ICM 20649, and could be a typo
There was a lot of activity around EKF2 lately. Could you provide a birds eye view of problems and solutions related to that? Thanks!
I meant Github checkins in Main of course
- Some logging changes
- Bug fixes
- Yaw estimator improvements
- Check out the release notes too: v1.14-RC1 Release note update by junwoo091400 · Pull Request #2645 · PX4/PX4-user_guide · GitHub
- Also check the PRs that were merged: https://github.com/PX4/PX4-Autopilot/pulls?q=is%3Apr+is%3Aclosed+label%3AEKF2
px4 messages repository getting updated via CI is time consuming, is there way to bypass this?
- In the future, we want to have the PX4 be built via colcon itself, so we don’t have to sync these message definitions
- Can replace message folder with symbolic link as workaround.
Having a FC with multiple locations of storage location use case.
- Configuring parameter storage location via parameters isn’t ideal
- We can rather place the symbol in the KConfig
- The location should be set at compile time, not run time, so KConfig would be better option?
- Currently parameters, waypoints, logging are ones we are discussing about.
- Peter can create draft PR on using KConfig instead
Are there plans to drop support for gazebo-classic? What is the long term plan for simulation?
- After Ubuntu 20.04, it will be discontinued, so we will have to make the new Gazebo as the main development
- After few tweaks, new Gazebo should be as usable / better as the classic Gazebo.
- Timeline wise, we won’t be dropping classic Gazebo any time soon, so you can just use the new Gazebo (it’s a choice)!
- Please try the new Gazebo, it’s fairly good → Gazebo Simulation | PX4 User Guide (main)
Is it possible to have wind estimate without airspeed sensor?
Currently it is assuming no side slip, and then tries to guess the wind speed. Should work only when plane is doing quite a lot of turns and wind is somewhat constant.
Having a new module written as child class of WorkItem, instead of ScheduledWorkItem, results in a module when being stopped (via NSH), to not exit properly
Also, having the
bool init() function defined in the header file of the module class results in error:
error: bit-field ‘bool MyModule::init()’ has non-integral type ‘bool (MyModule::)()’"
So it seems like there may be some issues in how I constructed the module, but no clue for now
Currently the requirement is quite high end, is there way to run it on low power environment / laptop? Would like to test on 4-core CPU.
- This is quite NVIDIA specific, but still could be interesting to get it supported
- In Academia people are utilizing this simulator more nowadays for Reinforcement Learning (also for Hardware testing)
- Figuring out plugins for GPS, etc would be a consideration too.
- We could use this as a ROS2 bridge example as well
Action Item: Jay and Daniel will keep it in mind and get in touch with the simulator authors in the future.