Adaptive and Secure Autonomy for UAS: A Modular Approach to Flight Control, Machine Learning Integration, and Multi-Vehicle Coordination

This proposal outlines the SwiftCore Flight Management System, a modular avionics solution for uncrewed aircraft systems (UAS) designed for high reliability and rapid adaptability. SwiftCore utilizes a layered architecture with intelligent nodes communicating through a publisher-subscriber messaging system. Its purpose is to provide a flexible and safe platform for advanced UAS functionalities, including multi-vehicle coordination, machine learning-based sensing and control, and adaptive tuning. Phase I funding will be utilized to further develop and validate this modular, layered architecture. Key technical objectives to be achieved with the funding include: -Defining and documenting a modular, multi-layer control and sensing architecture with standardized messaging based on the SwiftCore which already contains many of these characteristics. -Adapting and refining the existing SwiftCore system to meet evolving operational requirements and support multiple aircraft configurations. -Developing a "safe sandbox" environment to allow testing of experimental control modules without compromising overall system safety. -Demonstrating a proof-of-concept for machine learning–enabled, in-flight tuning of control loops. The target markets for the SwiftCore system include government agencies like NASA, NOAA, USGS, and the DoD, as well as commercial customers in aerial surveying, agricultural monitoring, and infrastructure inspection. Uncrewed aircraft original equipment manufacturers (OEMs) and academic research institutions are considered important target market segments.