Modeling and Simulation of Vehicle Performance in a UAV Swarm Using Horizon Simulation Framework

Abstract

A UAV swarm is simulated using Horizon Simulation Framework. The asset utilized for the swarm agent is a simplified model of the MQ-1 Predator, a large fixed-wing aircraft. The simulated swarm utilizes a decentralized cooperative control approach to command the assets through the use of digital pheromones and a pheromone map. Each vehicle operates at steady-state flight conditions of 36 m/s with an altitude of 1,800 m, and utilize an LQR set-point controller to maneuver through the pheromone map. All pheromone and aircraft related models are written in Python to expand the HSF scripting capability and include airborne scenarios. The simulation study focuses in the variation of three parameters in the repelling pheromone model. The first two are the update and deposit parameters with values of 2, 10, and 18. The third is the threshold parameter with values of 1e-02, 1e-10, and 1e-18. The lower parameter values provide more time-on-target while the higher parameters allow the swarm to search the surrounding area by only visiting the grid-space once. iv ACKNOWLEDGMENTS This thesis is dedicated to all "Magnificent Bastards" from 2nd Battalion 4th Marines, and to all of my other 5th Marine Regiment brothers whom I served with. Thanks to Cpl Dustin Libby and SgtMaj Joseph Ellis for motivating and influencing me to never give up on my goals. Semper Fi brothers, you will never be forgotten. Thank you to Dr Paulo Iscold and Dr Colleen Kirk for taking time out of their busy schedule to be part of my committee. Thank you to Dr Kira Abercromby for making aerospace courses fun and more interesting; thanks for giving me the inspiration to pursue this topic. Thanks to all the faculty and staff of the Aerospace Department for all your passion and dedication to students. Thanks to my family and friends for their support. Thanks to my Hufflepuff friend Karen Kauffman for introducing me to some great reading material, and for all the riveting programming and hackathon conversations. Thanks to my life mentors and friends, Dr Richard Robertson and Craig Snow, for all their valuable advice, and a special thanks to Dr Eric Mehiel for being my advisor. Thank you for being a positive influence in my life and for introducing me to the world of dynamics and control that I have come to enjoy so very much. Thank you for all your guidance and patience with me during the writing of this thesis. v