Date

2019

Author

Yılmaz, İsmail Çağdaş

Metadata

Show full item record

Item Usage Stats

408
views

162
downloads

Cite This

This thesis covers parallel docking (parallel parking) problem for unmanned surface vehicles (USVs). First, a mathematical model for a USV with two propellers is constructed by using Newton-Euler formulation. Kinematics and dynamic equations create 6 degrees-of-freedom model. A hierarchical motion control approach is implemented on this model. Two kinds of guidance laws, line-of-sight (LOS), and pure pursuit (PP) are employed for way-point travelling at the strategic level of the hierarchy. At the control allocation level, a finite horizon model predictive controller (MPC) and a cascaded PID controller are designed and tuned to optimize path following the performance. These guidance and control methods are implemented for parallel docking, which is treated as a way-point generation problem. Path generation for docking is handled in two stages. In the first stage, by solving a constrained optimal control, a path is found which provides that the vehicle reaches the port of the parking region with minimum control demands. By using a continuous curvature path function, the vehicle is taken from port to parking slot. The path following and energy consumption performances of the USV under the parallel docking manoeuvres are evaluated for different combinations of guidance laws and controller designs at the second stage. Finally, experimental validation has been realized on a scaled boat with model predictive control and pure-pursuit guidance methods.

Subject Keywords

Vehicles, Remotely piloted., Mathematical models., Marine machinery., Propellers.

URI

http://etd.lib.metu.edu.tr/upload/12623149/index.pdf
https://hdl.handle.net/11511/28072

Collections

Graduate School of Natural and Applied Sciences, Thesis