Collective Inference Based Data Analytics System for Post Operations Analysis Phase II
Status: Completed
Start Date: 2018-05-16
End Date: 2020-10-24
Description: This SBIR research provides a significant improvement over current-day post operations analysis (POA) with significant commercialization potential. In Phase I, ATAC developed a machine learning based aviation POA decision support tool (DST), which improves the state of the art of today’s airline, airport and FAA POA processes by providing automated, results-oriented POA outcomes. We provided a proof-of-concept for this POA DST by demonstrating how the Phase I prototype allows airline, airport and FAA personnel at the Charlotte Douglas International Airport (CLT) to perform faster, more efficient and results-oriented post analysis of individual departure banks to obtain actionable operational insights. Encouraged by our promising proof-of-concept demonstrations in Phase I, we propose to carry forward this research in Phase II of our SBIR project towards the eventual goal of developing a commercial licensable Cloud-based POA Platform that can be accessed by NASA, FAA, airline, airport or other commercial systems or personnel in a “Platform-As-A-Service” (PAAS) mode. This proposed capability provides a one-stop platform for gate-to-gate, complete POA including aviation data acquisition, storage, analytics, and root cause diagnosis, in a post-analysis mode as well as a real-time, continuous operations monitoring mode. The proposed continuous operations monitoring mode accelerates operations analysis work related to NASA’s ATD-2 project. The proposed second airspace focused use case supports multiple NASA research programs, including ATD-2's CLT to Northeast corridor (NEC) departure flow operations analysis, IDM NEC enroute constraints analysis, ATD-3 weather-efficient routing analysis and System Wide Safety anomaly detection. Moreover, by providing the ability to perform results-oriented POA on diverse operations (UAM, IDO), the SBIR enables the future NAS to rapidly learn from operational inefficiencies, and improve new traffic management and operations paradigms.
Benefits: The CIDAS-P technology has application across multiple NASA projects. CIDAS-P's automated, results-oriented post operations analysis (POA) complements ongoing ATD-2 CLT operations analysis efforts. Phase II continuous ops monitoring and NEC surface-airspace ops analysis capabilities support development of ATD-2's Strategic Scheduling component, and accelerates progress towards the next phase of ATD-2. CIDAS-P collective inference analysis on weather-driven en route rerouting scenarios can accelerate evaluation and continuous improvement of ATD-3 rerouting technologies. ATM-X's IDM research will benefit by leveraging CIDAS-P as a reliable, results-oriented method for evaluating the effectiveness of enroute TBFM-TFMS coordination strategies. CIDAS-P's airline ops analysis use case supports NASA Airline Operations Research Group (AORG) in its objective of infusing NASA-funded technologies into airline tools. CIDAS-P can also significantly improve System Wide Safety (SWS) project's anomaly detection algorithms, by providing reliable, automated collective inference based guidance on whether the identified safety alerts are false positives or missed safety alerts. Research into future diverse operations (UAM, IDO) also stands to benefit by CIDAS-P enabled continuous operations improvement guidance. Working software prototypes and collective inference algorithms can be incorporated into NASA software analysis platforms such as DASH, SMART-NAS testbed, FACT, or FACET.
The main commercial application for the proposed technology is as a DST to be used by operational and/or analytical personnel at airlines, ANSPs, or airports, (or by aviation consultants) for analyzing root causes for observed operational efficiencies or irregularities at the end of a day of operations at key airports and airspaces. The CIDAS-P collective inference engine will enable staff to differentiate the impacts of factors under their control versus not under their control, to assist in improved operational decision-making around operational procedures, and technology and resource investments. Airline uses include better analysis of irregular operations responses, improved analysis of airline network-wide flight scheduling and management, fleet mix choices, gate turnaround, gate pushback, and non-movement area operations, diversions and cancellations, and competitive airline performance. In the case of airports, uses include the analysis of the impact of airport construction schedules, departure metering operations, and general management of gate turnaround, gate pushback and non-movement area operations. In addition, specific ANSP-focused applications include: (1) a Trajectory-based Operations (TBO) benefits analysis and monitoring capability, (2) an operational analysis tool focused on measuring the impact of ATM DSTs for departure metering, weather rerouting, and arrival metering, (3) NAS weather impact analysis tool, and (4) a post-operations TFM evaluation system.
The main commercial application for the proposed technology is as a DST to be used by operational and/or analytical personnel at airlines, ANSPs, or airports, (or by aviation consultants) for analyzing root causes for observed operational efficiencies or irregularities at the end of a day of operations at key airports and airspaces. The CIDAS-P collective inference engine will enable staff to differentiate the impacts of factors under their control versus not under their control, to assist in improved operational decision-making around operational procedures, and technology and resource investments. Airline uses include better analysis of irregular operations responses, improved analysis of airline network-wide flight scheduling and management, fleet mix choices, gate turnaround, gate pushback, and non-movement area operations, diversions and cancellations, and competitive airline performance. In the case of airports, uses include the analysis of the impact of airport construction schedules, departure metering operations, and general management of gate turnaround, gate pushback and non-movement area operations. In addition, specific ANSP-focused applications include: (1) a Trajectory-based Operations (TBO) benefits analysis and monitoring capability, (2) an operational analysis tool focused on measuring the impact of ATM DSTs for departure metering, weather rerouting, and arrival metering, (3) NAS weather impact analysis tool, and (4) a post-operations TFM evaluation system.
Lead Organization: ATAC