PITT-STOP-PRO

Breaking the limits of speed!

Every second counts in an f1 race. Currently drivers rely on instincts and often judge their best performance by comparing with previous track records.
This leads to unreliable assumptions about cornering (turning over close corners) and optimal speeds.
A control model wherein optimal speeds, force and turning angle (control inputs) can be obtained given a racetrack coordinate data for each small time step could replace the assumption based approach.
This would lead to better "tracking simulations" for drivers to improve their performance!

Our team has successfully developed an optimal control methodology wherein we first find a center line of any given racetrack in global coordinate frame.
Then we find the minimum curvature optimal racetrack line which does not involve vehicle dynamics in play.
Including vehicle dynamics in picture, we generate the optimal racetrack which would finally be traced by a low-level controller.
We hope that this project will help develop better training simulators for racecar drivers in order to reduce lap-time and optimize racecar performance as well.
Please check the paper below for detailed explanation on the methodology, mathematics and result discussions.

Designed, researched and developed a geometric QP to find the minimum curvature racetrack model.
Researched and initially implemented a low-level bicycle dynamics model for the IPOPT solver.
Worked on developing the low-level controller to track the given race line generated after taking vehicle dynamics in account.
Debugging the code and plotting the results.

OCRL_paper.pdf

Page updated

Google Sites

Report abuse