Embedded Control of a Rotary Inverted Pendulum

Abstract: In modern-day contemporary control systems, such as robots, drones, and autonomous vehicles the control algorithms are bound to be implemented on embedded devices that pose a set of challenges and limitations such as low computational capacity and memory. Thus, it is crucial to design and implement efficient control algorithms while using low-level programming languages such as C / C++. This thesis presents the implementation of an embedded interface between a classical control system testbed, specifically a rotary inverted pendulum, and a microcontroller board. Control algorithms such as the Kalman filter and optimal LQR controller are employed along with a wireless networked control loop and a FDI method that maintains stability under a potential sensor fault. The control algorithms are implemented and evaluated on a microcontroller board. Furthermore, an overview of potential embedded constraints that may arise while implementing the control algorithms on the microcontroller board is presented.