[alhirzel]

The way this is described in a control systems framework is that the dynamics are non-minimum phase. A cool high-level video from Mathworks[1] explains in more mathematical detail. Being an experienced rider but having not looked into bicycle models in detail, I think it would make intuitive sense if the dynamics were only NMP for some operating points, particularly very low roll rates at very low roll angles. Or equivalently, the system behaves differently when it's in a banked turn (moderate roll with low roll rates) because gravity is starting to couple into the steering more. There is an angle where the steering switches modes depending on the geometry of the bike, and I think this is why road bikes "feel" "twitcher".

[1]: https://www.youtube.com/watch?v=jGEkmDRsq_M

Edit - also OP, very cool thesis work!

[kevin_thibedeau]

Road bikes are twitchier because they have less trail. Even 25 year old MTBs are twitchy compared to modern equivalents succumbed to the fashion for negative stem lengths and slack head tubes trying to imitate DH bikes.

[alhirzel]

Good point and makes sense - probably the reduced trail dominates most of the feel. I wish I had time to dive into a model to see if there really is any non-minimum-phase behavior.

Also, I skimmed the following to help my understanding; others may find it useful too: https://www.cyclingtips.com/2018/11/the-geometry-of-bike-han...