Date of Award

3-2021

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Operational Sciences

First Advisor

Matthew J. Robbins, PhD

Abstract

Within visual range air combat requires rapid, sequential decision-making to survive and defeat the adversary. Fighter pilots spend years perfecting maneuvers for these types of engagements, yet the emergence of unmanned, autonomous vehicle technologies elicits a natural question - can an autonomous unmanned combat aerial vehicle (AUCAV) be imbued with the necessary artificial intelligence to perform air combat maneuvering tasks independently? We formulate and solve the air combat maneuvering problem to examine this question, developing a Markov decision process model to control an AUCAV seeking to destroy a maneuvering adversarial vehicle. An approximate dynamic programming (ADP) approach implementing neural network regression is used to attain high-quality maneuver policies for the AUCAV. ADP policies attain improved probabilities of kill among problem instances most representative of typical air intercept engagements. Maneuvers generated by the ADP policies are compared to basic fighter maneuvers and common aerobatic maneuvers. Results indicate that our proposed ADP solution approach produces policies that imitate known flying maneuvers.

AFIT Designator

AFIT-ENS-MS-21-M-152

DTIC Accession Number

AD1130933

Share

COinS