|M.Sc Student||Uri Nenner|
|Subject||Control of an Unmanned Motorcycle Robot|
|Department||Department of Civil and Environmental Engineering||Supervisors||Professor Linker Raphael|
|Professor Emeritus Gutman Per-Olof|
|Full Thesis text|
Motorcycles are a fascinating mean of transportation. Unstable by nature, they demand a consistent rider effort to keep them stable. Apart from the inherent instability, in the absence of slipping a motorcycle is subject to non-holonomic contact constraints, meaning that only longitudinal movement is feasible. These make the development of an autonomous or tele-operated motorcycle especially challenging and very few studies have been devoted to this topic.
In the present study, the linearized equations of motion for a motorcycle were derived in a thorough manner using the Newtonian approach rather than the Lagrangian approach. Although the Lagrange formulation is easier, it leads to loss of direct physical interpretation, whilst the Newtonian formulation is transparent. These equations contain two degrees of freedom, the steering and roll angles. A stabilizing robust controller was synthesized based on these equations, utilizing the QFT method. An experimental motorcycle robot prototype was built, and the control scheme was validated first through simulations for various forward velocities, while including uncertainty in the model parameters. Then the control scheme was implemented on the motorcycle prototype and experiments were conducted successfully.