Theses and Dissertations

Advisor

Luo, Chaomin

Committee Member

Jones, Bryan A.

Committee Member

Ball, John E.

Date of Degree

5-16-2025

Original embargo terms

Visible MSU Only 2 Years

Document Type

Graduate Thesis - Campus Access Only

Major

Electrical and Computer Engineering

Degree Name

Master of Science (M.S.)

College

James Worth Bagley College of Engineering

Department

Department of Electrical and Computer Engineering

Abstract

Digital twin technology can play a significant role in mobile robots’ navigation by providing a virtual representation of the physical environment, robots, and their interactions. This high detail simulation can allow efficient and accurate navigation in difficult scenarios while enabling cost effective robot solutions. In this research a digital twin-based framework is proposed to facilitate mobile robot navigation throughout partially known, static environments, while making use of the strengths of a centralized system. The virtual complement digital twin of a real-world environment is first generated using previously known details such as static obstacles, walls, and passageways. The framework utilizes an improved version of RRT*-Smart for path planning, where Proximal Policy Optimization based reinforcement learning is trained using numerous planning trials of the simulation, slowly updating the algorithm’s parameters to fit the specific environment. During runtime, the digital twin system constantly updates itself in real-time using robot sensor data, allowing a dynamic window approach-based local navigation algorithm to path each robot to their respective destinations as well as improving any future path generation. The overall system is validated through the use of both path planning comparison studies as well as real-world simulation studies of navigation through a warehouse.

Sponsorship (Optional)

This research was partially supported by the Mississippi Space Grant Consortium under NASA EPSCoR RID grant.

Download

Share

COinS