Date of Graduation

2015

Document Type

Thesis

Degree Type

MS

College

Statler College of Engineering and Mineral Resources

Department

Mechanical and Aerospace Engineering

Committee Chair

Yu Gu

Committee Co-Chair

John Christian

Committee Member

Jason Gross

Committee Member

Powsiri Klinkhachorn

Abstract

A classical problem for robotic navigation is how to efficiently navigate from one point to another and what to do if obstacles are encountered along the way. Many map based path planning algorithms attempt to solve this problem, all with varying levels of optimality and complexity. This work shows a review of selected algorithms, and two of these are selected for simulation and testing using a quadrotor unmanned aerial vehicle (UAV) in a dynamic indoor environment which requires replanning capabilities. The Dynamic A* algorithm, or simply D*, and the Probabilistic Roadmap method (PRM) are used in a scenario designed to test their respective functionality and usefulness with the goal of determining the better algorithm for flight testing given a partially known or changing environment.;The development of the quadrotor platform hardware is discussed as well as the associated software and capabilities. Both algorithms are redesigned to fit this specific application and display their respective planned and replanned paths in an intuitive and comparable manner. Simulation is performed and an obstacle is added to the map during the quadrotor motion, requiring a replanned path. Results are compared for both computed path length and computational intensity. Flight testing is performed in an indoor environment, and during the flight an obstacle is inserted into the flight path, requiring detection and replanning. Results are compared for computed path length and intuitively analyzed to compare optimality and complexity.

Download

Share

COinS