AI/ML Aided Aviation Sensors for Cognitive and Decision Optimization

In manned aviation environments – both military and commercial – AI support is being developed and researched for ground-based planning and operational decision support. Using AI in real-time with crews poses additional questions and issues. This research will provide the means to understand the measures and requirements to architect potential AI-Agent solutions before implementation. This feasibility study will model surrogate data required to make in-flight route recommendations based on a changing, real-time mission environment using threats, current aircraft positioning and mission objectives (time, location). The objective is to determine the compute environment, optimum machine learning architecture and, most importantly, the Human-Machine or Human-Agent Interaction (HMI/HAI) measures that will best determine crew trust in using an AI-agent in mission environment. In large data, or multiple data stream, environments have several available solutions to mitigate levels of human attention, cognition, and the volumes of information incoming. Methods such as data fusion can reduce the volume of incoming information, machine learning algorithms are capable of statistically learning how to classify or pattern recognition from fused or raw data, and tailored dashboards can provide analysis results and focusing user attention to critical findings. Each of these tools can transform how users interact with data; but when coupled, they can provide a foundational single source of human machine interaction (HMI) enabling individuals to navigate a multitude of unique data streams with unique meanings to support decision-making with approachable and useful information.