Imaginary Values Appearing in the Leg Chains

[wgtayar]

Hello,

First of all, this is such a great work, thanks for posting it!

I have been experimenting a bit with the code, and am trying to run long simulations (about 30mins long). However, I faced an issue that is hindering the completion of the calculations: if I make the SimTimeDuration variable larger than 1000, the algorithm fails to provide a real set of solutions for all chains of the legs.

Then I started investigating more into the issue, and I found that if I do not run clear all before running MAIN, meaning if I do keep the last values of the simulation, with p and all other variables as they were, the same outcome is observed: imaginary numbers start popping up in the chains of the legs.

I am skeptical of two things:

1. I am using quadprog and not qpSWIFT.
2. The fact that I am "reusing" p's entries may provoke unstable starts due to the last state the quadruped was in

Would you be able to support me on this issue? I would greatly appreciate it.

Thanks in advance!