search

alexliniger / MPCC

Model Predictive Contouring Controller (MPCC) for Autonomous Racing
Apache License 2.0
1.38k stars 373 forks source link

Regarding the addition of dVs to the control input #90

Closed kms8527 closed 9 months ago

kms8527 commented 9 months ago

image

Given the dynamics model, the control input dVs is used to increase \dot{s}, yet s seems to have no mathematical correlation with the vehicle's speed (vx, vy). Thus, I suspect that dVs and \dot{s} could not possibly affect the MPC performance, but in practice, it appears not to be the case. dVs is a virtual input that cannot be directly applied as a control input in real vehicles, and I'm unsure why Vs was introduced. Could you theoretically explain this?

alexliniger commented 9 months ago

The connection comes via the lag error which forces s to be close to the projection of X and Y onto the reference path. Given that s is linked to the position v_s and dv_s are also linked to the position.

The result is similar to Frenet/Curvilinear coordinates but while keeping the Euclidean position.

kms8527 commented 9 months ago

thank you

fzhang327 commented 3 weeks ago

Thank you for your question, I also notice this. Is this the arc-length dynamics described in "AMZ Driverless: The Full Autonomous Racing System"?

kms8527 commented 3 weeks ago

yeah, i think so. you can also check "Optimization-based autonomous racing of 1:43 scale RC cars".