Fix the ICSC 2023 figure

[moorepants]

With recent updates it no longer gives a stable simulation. It should look like what we had, see: #15 (comment)

[moorepants]

I went back to e6db13f and ran it. It gives a stable simulation.

I have matching parameters and gains in the current version (other than the tire compression parameters) but the simulation gives different results.

Current commit 2f1d766;

[moorepants]

It may be that the self aligning moment was incorrect or set to zero in the original abstract plot.

[moorepants]

I am thinking that the differences in the model are enough to change the result. You can tweak the gains a bit in the sim (even just using T7 = 10*u4) and get a similar behavior as the icsc original simulation. So maybe there is nothing to solve here other than picking new control gains to get similar motion as before.