Surrogate-Based Optimization for Radar Surveillance Requirements to Support RWC and CA for Unmanned Aircraft

The integration of unmanned aircraft into the existing airspace will require reliable automated sensors and systems to ensure safe separation and operationally acceptable guidance. To support this goal, the aviation community is currently developing standards for Detect and Avoid (DAA) systems. In parallel, the Airborne Collision Avoidance System Xu (ACAS Xu) is under development by the Federal Aviation Administration (FAA) as a DAA implementation that provides both Remain-Well-Clear (RWC) and Collision Avoidance (CA) alerting and guidance. The development of requirements for surveillance sensors and systems to support the required levels of safety and operational performance defined by the aviation community necessitate extensive parametric studies to evaluate end-to-end performance across a wide range of potential designs. With a high-dimensional design space, this task becomes intractable due to the computational cost of simulating a large set of encounters for each potential set of design parameters. Additionally, performance of ACAS Xu must be examined in the context of two alerting schemes: CA only and with both CA and RWC alerting. In this paper, we demonstrate the use of surrogate optimization techniques to derive surveillance performance requirements to support both RWC and CA alerting with ACAS Xu. We will illustrate the robustness of the framework for efficiently deriving requirements by performing the process both with and without RWC alerting. Additionally, we will show the ability to incorporate vendor design preferences and additional constraints on the design space imposed by the drafted DAA requirements. The results will substantiate a repeatable, flexible framework for deriving radar surveillance requirements, supporting the development of the ACAS Xu Minimum Operational Performance Standards (MOPS) and harmonizing the standards for DAA and CA.