Quantifying AAM Communications Quality using Machine Learning

Abstract

Achieving Advanced Air Mobility (AAM) on a scale envisioned by industry proponents and other stakeholders will require an Air-Ground Communication (AG Comm) system that is robust and resilient to failures. In this paper we describe a Machine Learning-based tool that quickly predicts communication path loss for AAM flights, a key metric for establishing and maintaining robust AG Comm. We have implemented this tool and tested it using both simulated scenarios and live flight data. This paper describes the tool itself and the results obtained comparing it with "ground truth" as established through physics-based ray-tracing computations.