Analyzing Identification Friend or Foe (IFF) of Armored Vehicles: A Comparative Approach with Transfer Learning and Pre-Trained Models

In military operations, the proper use of force in accordance with the laws of war are non-negotiable. As the battlespace is becoming more and more volatile, uncertain, complex, and ambiguous, military leaders are becoming more dependent upon advanced technologies such as artificial intelligence to assist in their decision making. One such area that military leaders need advanced technology support is positive control of all friendly assets as well as identification of enemy and noncombatant assets in all domains of air, land, sea, and space through the identification of friend or foe (IFF). The purpose of this study is to assess the viability of deep transfer learning to assist in military decision-making and answer the research question: Can pre-trained deep learning models be used to adequately identify and classify enemy targets? This study is designed as a comparative study to assess and compare various pre-trained deep-learning models to determine if they are adequate for targeting and engaging the enemy. An ensemble model was also incorporated using three pretrained models and compared to the results of the individual models. A discussion of Human-in-the-loop concepts as well as the ethical considerations of the use of AI for IFF is incorporated in this study.