Bachelor and Master Theses

In the context of the Mälardalen University Automation Research Center (MARC) research profile, a use case is set up to support research related to digital infrastructure and the computing continuum. The system consists of distributed nodes for control and image processing, communicating with the pendulum setup. In this thesis, 2-3 different high-level control strategies for distributed Furuta pendulum swing-up and stabilisation will be selected and implemented. The stability and performance of the strategies will be evaluated in the presence of different amount of latency and jitter in sensor data arrivals to the high-level control component, and in the communication from the high-level control to the low-level control.   

The low-level control and sensor retrieval will be implemented in C/ C++ on a Raspberry Pi connected to a Quanser QUBE Servo with an inverted pendulum module.  The high-level controllers will initially be evaluated on the same node, but as a separate module that may be moved to another node on the (TSN) network. If one or more strategies can tolerate sufficient latency in the sensor data, it will be moved to another node and/ or integrated with a camera-based retrieval of the pendulum state.

• Control theory 
• Knowledge of linux and C/ C++. MATLAB/ Simulink may potentially be used for part of the work

Quanser QUBE Servo: https://www.quanser.com/products/qube-servo-3/