Kenneth Pasma successfully defended "Model Matching Control Applied to Bicycles" on August 28, 2024. Kenneth developed a controller for our steer-by-wire bicycle's steering motor that enables matching the bicycle's dynamics to that of a reference bicycle. A model matching controller strives to make the input-output dynamics of the controlled vehicle the same as a reference or target vehicle (or vehicle design). In essence, this means that we can make a steer motor-controlled bicycle feel and move as if were another bike, for example one with a different trail, or different wheel base, or with less mass. Kenneth applied the approach of applying both feedforward and feedback control to the steering motor based on force and motion measurements from sensors on the bike. Once the controller was functioning, he performed two experiments to prove whether his controller was working as expected: 1) identification of the weave mode from a lateral pulsive perturbation and 2) identification of steer torque to steer and roll kinematic transfer functions through sinusoidal forcing of steer torque. Kenneth's results show that the controller shifts the weave mode's decay and frequency behavior towards that of the reference bicycle and that the input-output magnitude relationships at different forced frequencies also shift toward that expected of the reference bicycle. Kenneth explained how these techniques could be used for virtually testing different bicycle designs and how it can open possibilities for new experiments in bicycle handling and control. We also hope to use the technology to make our balance assist bicycle's feel more natural to the rider while retaining the stability effects.

Kenneth was co-supervised by Prof. Thomas Kavicsky (TU Delft) and Jason K. Moore (TU Delft). Everyone at the bicycle lab is very proud of Kenneth and wishes him the best in his next adventures.