Overview
What this challenge is about.
You receive a kinematic bicycle model of the robot, a published track layout, and 30 minutes of recorded waypoint trajectories. Implement a nonlinear MPC controller using acados or CasADi with a 1-second horizon and 50ms control frequency. Compare against the existing PID controller on: cross-track error, control effort, and behavior under simulated wind disturbance. Recommend whether MPC should ship to the next on-sidewalk pilot in Zurich's Old Town.
The Brief
What you'll do, and what you'll demonstrate.
Determine whether nonlinear MPC outperforms the PID baseline enough to justify a controller swap in production.
Earning criteria — what you'll demonstrate
- Implement nonlinear MPC for a real robotics platform
- Compare classical (PID) and modern (MPC) controllers on equal footing
- Analyze controller robustness to disturbances
- Document a controller decision the chief engineer can defend
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
AI Engineer
Owning a controller swap with a defensible decision document is exactly the kind of project that lets a junior AI engineer earn ownership of a real robotics subsystem.
This challenge sharpens
- model-predictive-control
- control-systems
- python
Machine Learning Engineer
Implementing optimization-based controllers and benchmarking against incumbents is a direct skill transfer to MLE roles on learned-controller teams.
This challenge sharpens
- model-predictive-control
- optimal-control
- trajectory-tracking