Doctor of Philosophy (PhD)
Computer Science and Engineering
Robots are increasingly utilized to perform tasks in today's world. This has varied from vacuuming to building advanced structures. With robots being used for tasks such as these, new challenges are introduced. Problems that have been previously researched to be performed, either theoretically or implemented, need to be redesigned to be able to better handle these challenges. In this thesis, I will discuss multiple problems that have previously been researched and I have redesigned to be possible to be implemented by robots or that I have developed a new way for the robots to solve the problem. I focus on geometric areas and robots tasked with performing exploration in the area. Exploration is a task in which an unknown area is completely traversed. In this work, I have develop multiple algorithms to perform online tasks that can be implemented by robots. With robots performing exploration, a limited viewing range and communication range increase difficulty. These algorithms are focused on utilizing robots to perform Exploration and Concave Decomposition. The results from this thesis are such that the Exploration algorithm that given a fleet of robots k, the total area n can be explored in O(n/k) time with all agents having work O(n/k). The Concave decomposition task has multiple algorithms focusing on a different aspect. In the first, with an online algorithm, with r reflex points, I perform the decomposition generating r + 1 convex areas. The other two online algorithms focus on interior cut length, which previously has not been researched. In response, have developed an algorithm which maintains interior length relative to the perimeter.
Clements, Wyatt Preston, "Geometric Problems in Robot Exploration" (2019). LSU Doctoral Dissertations. 4930.
Available for download on Saturday, May 16, 2020