Go-Around Prediction Service, Phase II

Our proposed innovation is a Go-Around Prediction (GAP) service that encapsulates predictive analytics so that stakeholders of NASAs In-time System wide Safety Assurance (ISSA) strategic thrust can readily use it to assess the go-around probability in real time during aircraft arrival operations in the National Airspace System (NAS). Our innovation is directly relevant to Subtopic A3.03 Future Aviation System Safety and fills two critical gaps in the state-of-the-art. First, it allows for the continuous monitoring of the NAS arrival domain and fuses diverse data sets including airborne trajectory, surface tracking, and weather data to identify the precursors to a key indicator of risk in the system (i.e., a go-around). Second, it applies innovative machine learning (ML) techniques to build and train models using historic go-around occurrences to predict go-around safety margins in real time. A key outcome in the first decade of ISSA-related research is improved safety through initial real-time detection and alerting of hazards at the domain level and decision support for limited operations. Our innovation directly addresses this outcome by focusing on the near-airport (within 10 miles) domain to identify risks to stakeholders (e.g., air traffic controllers and pilots) in enough time (before a go-around is necessary) for them to employ effective risk mitigations. Through the combination of a real-time data input stream and a ML based predictive model, the software service allows for the continuous computation of the go-around probability. Through integration with NASA platforms such as the In-time Aviation Safety Management System (IASMS), results can be updated and displayed to operators (i.e., air traffic controllers and pilots) for each arrival flight. This information will provide operators increased situational awareness during the approach phase of flight leading to earlier mitigation of developing risks and more time to safely manage go-arounds. Go-arounds or missed approaches are a clear risk indicator in NAS arrival operations. Occurring thousands of times per year across major airports, they are the mitigation of last resort for operators causing disruptions to the efficiency at the arrival airport from an air traffic control and airline perspective. They increase the workload for controllers and pilots when workloads are peaked, and the situation can become less orderly and predictable. Our innovation fuses real-time data feeds and applies innovative machine learning to continuously monitor arrival operations and reliably predict probabilities of go-around occurrence and their safety risks, thus addressing key gaps identified in Subtopic A3.03. A key innovation is that we go beyond static prediction at a given approach point and provide continuously updated go-around predictions as the flight progresses to touchdown. This capability supports in-time safety assurance to avoid go-arounds or improve their safe and orderly execution by providing more time for operators to act, thus increasing operational safety margins. The overarching objective of Phase II is to advance the proof-of-concept of the Go-Around Prediction (GAP) service developed in Phase I by maturing the system so that it provides value to aviation stakeholders through integration with NASA’s In-time Aviation Safety Management System (IASMS) and other platforms. The resulting prototype GAP service will demonstrate the ability to reliably predict the likelihood of go-around in advance and provide alerts to stakeholders. Our work plan includes: (1) Refinement of a clear requirements definition from a safety and value proposition perspective through interviews with airport, airline, FAA, and NASA SMEs. (2) Enhancement of the GAP service to include more operational scenarios such as mixed-use runways, complex arrival procedures, additional airports, and applicability to historical incidents/accidents. (3) Validation and enhancement of machine learning models for go-around prediction focusing on improving accuracy and reducing false positives. (4) Integration of the GAP service into NASA’s IASMS and Digital Information Platform (DIP), as well as ATAC’s commercial SkyView Data Services (SDS) architecture to provide access to stakeholders. (5) Pursuit of Commercialization Opportunities. Deliverables to NASA include a kickoff briefing, quarterly progress reports, a final report as well as prototype demonstrations of the GAP capability.
Benefits: 1. GAP advances NASA SWS research by accelerating risk detection to real-time. 2. GAP integrates with the In-Time Aviation Safety Management System (IASMS) to assess operational safety and identify emerging risks potentially introduced by new DSTs during initial deployment. 3. Integration with NASA’s Digital Information Platform (DIP) provides valuable go-around information to aviation stakeholders. 4. The predictive analytics service serves as a model for other NASA ARMD analytics developments. 1. ANSP personnel use the GAP capability to identify risks in airport operations much sooner than currently possible, thereby increasing the safety margin. 2. Airlines and airports use GAP to provide insight into go-around causes with the intent of reducing their risky and disruptive nature at major airports. 3. Automated Safety Management System (SMS) reporting for ANSPs, airlines, and airports