I am currently working on the flight computer for this https://arc.aiaa.org/doi/pdf/10.2514/6.2016-5532 as my master thesis. I plan to use UAVCAN on 5 nodes. One Linux based “Master” node and 4 nodes on the ATmega familiy (S128, maybe also S64M1).
For the ATmegaS128 I use an MCP2518 as a CAN controller and if I use the S64M1 it will be the in build CAN controller. The Linux maschine will be from the NVIDIA Jeston familiy with build in CAN FD controllers. Transceiver will be an LTC2875 on all nodes.
So technically I could at least drive 2 nodes on CAN FD (S128, Linux).
This will require the Libcanard and Libuavcan/Pyuavcan.
At the moment I am testing around with the driver and inter connectivity on the S128. Since Libcanard has CAN FD only in a WIP branch, I am at the moment focusing on using it with normal CAN.
So at the moment I do not write this because I have an error. Rather I think it is a nice project that you might find interesting.
But I am always open for questions and also suggestions!
@pavel.kirienko I am the guy from github with the ISS story. Do you have any suggestions?
Thank you, this is super interesting. We would love to learn more about the architecture of your system and its data flows, if possible (no pressure though).
At the moment, the best way to go ahead is to use PyUAVCAN, since it’s already usable. As I said on GitHub, I should start with Libcanard as soon as my current work on the Specification is finished. Meanwhile, we could use help with PyUAVCAN (and other projects):
I think it would help us greatly if you could prepare a short write-up explaining the objectives of the system and how exactly UAVCAN was leveraged to help attain them. Anything that relates to UAVCAN (especially the things I mentioned above) is of interest.