IEEE LARS/SBR 2021 - Cooperative and Decentralized Decision-Making for Loyal Wingman UAVs

Authors: Luiz Giacomossi Jr., Stiven Schwanz Dias, José Fernando Bracalion, and Marcos R. O. A. Maximo. Presentation for IEEE LARS and SBR (Latin American Robotics Symposium/Brazilian Robotics Symposium) 2021. With the application of decision-making techniques, it is possible to automate many of the intelligent decisions of UAVs, leaving human commanders to focus on higher-level decisions in modern warfare scenarios, providing a crucial tactical advantage. This paper applies key methods to enable decentralized decision-making for autonomous combat UAVs, also known as loyal wingman (LW) drones. A simulator was developed to simulate to implement and analyze the behaviors of the so-called LW. We used as the main decision-making technique the construction of a behavior tree (BT) capable of providing decisions in a decentralized manner, and for parameters optimization, the PSO optimization algorithm was used. Our approach shows a promising result of a kamikaze elimination efficiency of approximately 93%.