Modified Depth-First Search for the Automated Design of RNAV Approach Procedures

Abstract

Instrument flight procedures (IFPs) are a cornerstone of the National Airspace System (NAS) and provide paths and instructions for aircraft to safely operate in restricted visibility and in congested airspace. Design of IFPs that adhere to the criteria developed by the FAA to ensure safe flight can be a tedious, iterative task. This is particularly true in areas with complex obstacle/terrain environments. Previous work was conducted by MITRE to develop criteria modules (CM) capable of analyzing a procedure design’s compliance with FAA criteria. In this paper, these engines are leveraged to evaluate batches of candidate procedures. To ensure convergence to an acceptable solution in a reasonable period, a modified depth-first search algorithm was designed to mimic the general design flow used by human procedure designers. By combining this algorithm with the capability of the CMs and a batch cluster for analyzing multiple procedures in parallel, it is shown that procedures can be built in tightly constrained situations by adapting common algorithms to the specific requirements of IFP design.