Artificial Intelligence to Improve Situation Awareness for AEW&C Operators

Abstract

In the Airborne Early Warning & Control (AEW&C) system, the operator needs to be vigilant, perform multiple tasks, communicate, and simultaneously handle different sources of information during a mission. The operator's main objective is to receive, interpret and distribute information and data provided by the AEW&C system to create a Recognized Air and Sea picture (RASP). The information is then used to, for example, perform fighter jet control and to detect anomalies. Increased capabilities of Command & Control (C2) systems and new threats in the environment make the operators’ tasks more extensive, complicated, and complex than ever before, which could potentially impair their Situation Awareness (SA) and in turn their decision making. The domain is associated with high risks and potentially devastating impacts of faulty decisions. Therefore, the importance of finding new technologies to facilitate SA cannot be neglected. The increasing technological sophistication regarding the rapidly growing adoption of Artificial Intelligence (AI) offers new possibilities for creating systems facilitating SA. This motivates conducting studies to explore and analyse the opportunities AI provides to facilitate SA for AEW&C operators, in particular critical roles as the Fighter Controller (FC) and Surveillance Operator (SO). This study aims to answer the questions of when and how AI could be implemented to facilitate SA. The questions were answered using a process adapting the research through design approach. The when-question was answered through system analysis, user studies, concept development and evaluations and the how-question were answered by creating guidelines through literature studies and affinity diagrams and testing the guidelines by applying them in the development and evaluation of concepts. The system was divided into four separate subsystems where goals, subjects, tools, and outcomes were specified. Seven categories of SA related challenges were identified for FC and six for the SO respectively. Four concepts were created for each role aiming to improve SA, where concepts of a talk-translation tool, a formation recognition tool, abnormality detection tool, and timeline tool were considered the most promising for further work. Guidelines for three phases in development of AI functions were developed: planning, designing, and evaluating. The guidelines confirmed the utility of design through research by being used to explore opportunities to implement AI, ideate and create concepts and evaluate them in an AEW&C context.