Feedback Control of Under-actuated Systems using Time-Varying LQR
This project discusses the usage of time-varying linear quadratic regulator for underactuated systems. A two-wheeled cart-pole system (Segway) is used as a reference to study the performance of this control method and compare it with other control methods such as state feedback and LQR. Dynamic model for this system is designed using non-Holonomic constraints. A total of six states of this system are controlled using these control methods. An optimized trajectory is generated for all the states and inputs using optimized based control and is used for time-varying LQR. Overall, the better performance is observed with time varying LQR compared to other methods at various initial conditions. An attempt to simulate these control schemas is done on gazebo simulator as well.
Report for this project can be downloaded here.
Instructor: Siavash Farzan