Motion Planning Using Supervisory Discrete Event Control
Open Access
Author:
Ukah, Ucheoma N
Area of Honors:
Computer Science
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Minghui Zhu, Thesis Supervisor Dr. John Joseph Hannan, Thesis Honors Advisor
Keywords:
Motion Planning Path Planning Robotics RRG
Abstract:
The purpose of this research was to synthesize supervisors for robotic motion planning. The supervisory algorithm presented in this paper combines the low-complexity of random sampling using the RRG algorithm and the theory of discrete event control to produce a plan that enables a robot to traverse to a goal while avoiding all obstacles. The proposed algorithm acts as a supervisor by selecting the subset of the input set that provides the highest reward i.e., greatest likelihood of reaching the goal state while at the same time maintains a high probability of staying clear of the obstacles. The aforementioned supervision is made possible through the formation of the randomly sampled graph (RRG), which provides data about each state in the environment. The proposed algorithm acts as a tool for autonomous machines, allowing them to self navigate while avoiding obstacles in uncertain environments in an anytime fashion.
