Date of Award
Master of Science in Systems Engineering
Department of Systems Engineering and Management
John M. Colombi, PhD
This research was driven by the increase of autonomous systems in the current millennium and the challenging nature of testing and evaluating their performance. A review of the current literature revealed proposed methods for verifying autonomous systems, but few implementations. It exposed several gaps in the current verification and validation methods and suggested goals for filling them. Through the use of modeling, software in the loop (SITL), and flight test, this research verified an autonomous swarming algorithm for unmanned aerial systems (UAS) and validated an exemplar of a testing framework. Thirteen sets of three-vehicle swarm data produced over two days of flight testing provided a baseline algorithm analysis. During these tests, vehicle separation distances deviated an average of 5.61 meters from the ideal state, with separation distance violations < 6:39% of the time. The swarm achieved a 0.27 m average deviation and 0.43% violation in the best cases. Average packet loss between vehicles was 4.94% at a 5 Hz update rate, with an optimal communication lag < 0:04 seconds. The multi-faceted empirical analysis created through the pairing of qualitative and quantitative analysis provided a complete understanding of vehicle behavior. This analysis also identified various areas of improvement for the algorithm and testing framework. The outcomes of this research formed a baseline testing continuum that is adaptable to a variety of follow-on investigations into formal verification of autonomous systems.
DTIC Accession Number
Combs, Corey A., "Verification of Autonomous Systems: Developmental Test and Evaluation of an Autonomous UAS Swarming Algorithm Combining Simulation, Formulation and Live Flight" (2019). Theses and Dissertations. 2329.