[Discussion] Integrated companion/fc pulled out #2250
Conversation
* New translations land_detector.md (German) * New translations land_detector.md (Japanese) * New translations land_detector.md (Korean) * New translations land_detector.md (Russian) * New translations land_detector.md (Turkish) * New translations land_detector.md (Chinese Simplified) * New translations pixhawk6x.md (German) * New translations pixhawk6x.md (Japanese) * New translations pixhawk6x.md (Korean) * New translations pixhawk6x.md (Russian) * New translations pixhawk6x.md (Turkish) * New translations pixhawk6x.md (Chinese Simplified)
Auterion changed where the flash_emmc.sh file was located. I have updated the read me to show the correct location. \home\px4vision\catkin_ws\src\px4vision_ros\tools\flash_emmc.sh
Fix dependancy list and typos
* New translations video_streaming_wifi_broadcast.md (German) * New translations px4_vision_kit.md (Japanese) * New translations video_streaming_wifi_broadcast.md (Japanese) * New translations px4_vision_kit.md (Korean) * New translations video_streaming_wifi_broadcast.md (Korean) * New translations px4_vision_kit.md (Russian) * New translations video_streaming_wifi_broadcast.md (Russian) * New translations px4_vision_kit.md (Turkish) * New translations video_streaming_wifi_broadcast.md (Turkish) * New translations px4_vision_kit.md (Chinese Simplified) * New translations video_streaming_wifi_broadcast.md (Chinese Simplified)
From #2231 (comment) - the file is created by default on systems that support it.
|
|
||
| The companion computer and flight controller must be set up to communicate with each other (typically using MAVLink and ROS2, over Ethernet or a serial port), and may also be set up to communicate with ground controllers, cloud and other systems. | ||
|
|
||
| The setup depends on the companion computer and its operating system, and how the companion and flight controller are connected. |
There was a problem hiding this comment.
This is still a bit further out but ideally we'll standardize this as well, so there's a common way to update and run components (e.g. avoidance or a custom ros flight mode).
There was a problem hiding this comment.
Certainly it would be good if the PX4 project had a recommended companion OS and update method for that OS. That would make it possible to have turnkey instructions and CI testing that would work for 80% of the world.
Everyone seems to want their own OS though, so perhaps all that would need to be containerized?
Anyway, not my problem to solve.
hamishwillee
changed the title
|
Thanks @bkueng . PX4 hasn't really "committed" to the companion software stack; my thinking was that this is not needed yet, but if I do this it might help encourage adoption? Make the hardware accessible and the software standardization will come? Option 2 is also good, and long term better. I'll leave this open in draft and perhaps look at this again.
Well, it's custom PX4 build with a sandboxed proprietary container for software development. So you might argue that it isn't "PX4" at all from a development point of view. |
|
FYI @julianoes , just added you in case you had interest/thoughts on companion computer structuring. I'm also leaning towards "option 2" in the description. |
You can use it with upstream PX4, and some people do. But yes the companion is currently largely proprietary. |
@bkueng Following our discussion on "longer term having a section on integrated FC/Companion" I thought I'd have a look at what it might look like. Essentially this makes the integrated fc/companion a peer of autopilots and cross links to the companion section for the software setup.
I am not 100% convinced of this; after all, there are other solutions that are "advanced", like running PX4 on a companion. Also it gets complicated if we start having non-Pixhawk FCs in these things.
Another possibility is to do (something like) this:
ie. moving FC and FC/Companion as the top level and having peripherals etc underneath. I don't think we need to consider this right now.