Optimal and Robust Control
B3M35ORR + BE3M35ORR + BE3M35ORC
This course is part of an already archived semester and is therefore read-only.
Model predictive control for SK80 wheeled bipedal robot
Completion requirements
Design and implement a model predictive controller for the SK8O robot. The controller will use a two-wheel inverse pendulum model of the robot (hips are locked in a fixed position, only wheels are controlled). Take into account maximum torque limits on the wheels. Choose whichever design tool you prefer (MPC toolbox in Matlab, Casadi in Python/Matlab, OSQP in Python/Matlab, ...), but have in mind that automatic code generation feature would be beneficial as it would facilitate deploying MPC to the real robot. The model predictive controller must be able to run on the Odroid N2+ computer in real time (at least 100 Hz).

Last modified: Wednesday, 17 February 2021, 12:27 AM