PhD thesis defense to be held on June 13, 2023, at 11:30 (New ECE Building, Room B.0.13)
Picture Credit: Eleftherios Loghis
Thesis title: Robust Control Applied to Surface Vessel Guidance
Abstract: This PHD thesis deals with the electronics design, modeling, and control of a surface vessel created for reaching specified sets of targets on a two-dimensional map. It is based on a remote-control model of a ferry boat ship, that was assembled by NTUA. All the electronic equipment used, apart from the remote-control unit and its receiver, were designed and manufactured by the author of this thesis. The details of the development are reported on the thesis, along with the methodology followed for its modelling and control. After the construction of the vessel, data for its response were collected by performing manual maneuvers using the radio control unit. This data was used to train a neural network, to predict the vessel’s response to thrust and steering controls, in a black-box modelling approach. The model that resulted was used to conduct simulations on the control methods developed for the vessel. For enabling the vessel to reach specified targets on the map, two types of supervisory controllers were synthesized. These algorithms monitor the position of the vessel in relation to its target and generate translational and angular velocity directives that serve as setpoints for the respective control loops. To effectively cope with the non-linearities of the system, a methodology based on the neural network model was developed, that also decoupled velocity and heading control loops. Robust control techniques were also employed, to reduce the effect of model prediction and control action quantization inaccuracies and attenuate the resulting error. The response of the vessel was then simulated using several test cases, to examine its maneuvering abilities. A comparison of the two supervisory controllers was also conducted. Finally, a mini swarm of three vessels was simulated, to investigate the expansion of the methods described in the thesis, in sets of related vessels.
Supervisor: Professor Emeritus Tryfon Koussiouris
PhD Student: Eleftherios Loghis