strange behaviour when running mpc example about fast_and_efficient

[yxyang]

Not exactly sure why this is happening. I just did a fresh-clone of the repo on Python 3.8.10 + Ubuntu 20.04 and the code runs fine.

It seems that there is some issue with the base angular velocity, which is interpreted as list[int] but should be list[float]. The angular velocity should always be a list[float] though. Maybe try casting it as float and give another try?

[stiefen1]

I've tried to cast it but it seems that no matter what I provide to the function (even a constant float list : [0., 0., 0.]) I got the same error. Could it be the version of any library I'm using ? Because when I'm installing the libraries I got an error about the version of numpy which is not compatible with that of tensorflow (I don't need tensorflow since I only want to use your mpc controller).

ERROR: tensorflow 2.12.0 has requirement numpy<1.24,>=1.22, but you'll have numpy 1.24.2 which is incompatible.

Thanks a lot,
Stephen

[yxyang]

Doesn't look like a numpy issue to me --- I tested the MPC with numpy 1.24.2 and it works fine as well.

It is a really weird error that I have not seen before... Can you double check that you have not modified the code in any unintended ways, and re-compile the C++ binaries?

If it still doesn't work, you can try the following python implementation of the same MPC algorithm. I recently wrote it, but have not tested it thoroughly yet. It should follow a similar interface compared to the current repo.

mpc_python.zip

[stiefen1]

Hello,

I've simply recompiled the binaries and for some reasons now it works ! Thanks a lot for your help!

Best,
Stephen