NAS Metering Impact Prediction and Collaborative Scheduling System

The proposed SBIR applies innovative air traffic data processing, machine learning (ML), and time-based scheduling emulation methods to develop a NAS Metering Impact Prediction and Collaborative Scheduling (NMIPACS) Service. NMIPACS is highly relevant to Subtopic A3.01 because it addresses a key ATM challenge related to improving efficiency for the near-future (2025-2030) NAS. A FAA-NASA-Industry SWIFT working group identified the lack of an early TBFM delay impact prediction capability as a top priority because it prevents the FAA and flight operators from collaboratively negotiating TBFM program parameters. This results in avoidable and inequitably distributed TBFM delays. To solve this problem NMIPACS develops the following innovative microservices (and the associated digital assets): (1) A TBFM SWIM Data Interpreter to parse the existing TBFM scheduling configuration, parameters, and scheduling times; (2) An ML-based Prediction Engine for predicting Estimated Times of Arrival (ETAs) at TBFM scheduling points (a key digital asset); (3) A TBFM Scheduling Emulation that applies the TBFM scheduling steps to the scheduling data, and computes an accurate estimate of the TBFM Scheduled Times of Arrival (STAs) and the per-flight delay impact of those STAs (another digital asset). We make the TBFM Scheduling Emulation flexible so that it can compute STAs for multiple user-specified TBFM parameter choices. This enables NMIPACS to operate in a real-time what-if analysis mode so that the FAA and airlines can use it to collaboratively evaluate strategies for managing TBFM delays or distributing them more equitably among flights. In Phase II, we expand NMIPACS to address multiple other applications that address TFMS-TBFM interaction issues, among other things. Thus, NMIPACS enables multiple advances to the way TBO is executed in todayrsquo;s operations and enables TBFM to advance from an open loop system to a higher performance one that includes the airlines in a closed loop.