State Space Exploration (MS)
At the MPC lab we have developed a learning MPC for iterative task. The controller is able to improve its performance by learning from experience. We have shown that the controller improves its performance until it converges to a steady state solution. The goal of this project is to investigate the optimality of the solution and to propose techniques to improve the performance of the system when it converges to a local minima of the problem.
Prerequisite: Solid control and optimization background.