Your browser does not support JavaScript!

Home    Collections    Type of Work    Graduate theses  

Graduate theses

Current Record: 27 of 1622

Back to Results Previous page
Next page
Add to Basket
[Add to Basket]
Title Retrainable Object for Ocean Sailing
Author Καραβέλας, Παναγιώτης
Thesis advisor Τσιρώνης, Γεώργιος
Reviewer Μπαρμπαρής, Γεώργιος
Abstract Autonomous cars are starting to become a part of our life after their first appearance in 1980s. Since then, several attempts have been tried both to improve them, but also to create other types of autonomous vehicles like planes, drones and boats. In this work we create a new version of an autonomous sailing boat. Where instead of any path finding algorithms we use a basic reinforcement learning algorithm called Q-Learning. We create two versions of an environment where is a sailing boat in the sea with no obstacles. Under this approach we tried different types of rewards in order to find the most appropriate. After all the simulations, we apply the trained model to a prototype of a sailing boat which was built in the University. For the controller of the sailing boat we use low-voltage electronic devises like Arduino, Esp32, several types of sensors and servo motors. As expected the results of the simulation reveal that going along the direction of the wind is much easier than going against it.
Language English
Subject Q-Learning
R.O.F.O.S.
Reinforcement learning
Issue date 2023-11-22
Collection   School/Department--School of Sciences and Engineering--Department of Physics--Graduate theses
  Type of Work--Graduate theses
Permanent Link https://elocus.lib.uoc.gr//dlib/0/0/c/metadata-dlib-1696916773-928903-12679.tkl Bookmark and Share
Views 209

Digital Documents
No preview available

Download document
View document
Views : 4