Collaboratively Innovating Safety Today for the Safe Air Transportation System of Tomorrow

The U.S. national airspace system (NAS) is set to begin the next phase of airspace modernization. Motivated by the need to improve system reliability as well as to ensure that the airspace can safely accommodate increased demand for mobility, sustainability, and economic growth, modernization is critical to the future of U.S. aviation. By leveraging new state-of-the-art facilities and technologies, and by building off NextGen and trajectory-based operations, tomorrow’s NAS will be more integrated, digitally connected, service-oriented, and performance-based.

While passengers and shippers are the beneficiaries of these expanded capabilities, airline pilots will help make certain that modernization results in a safe, secure, and reliable transformed NAS. NASA’s System-Wide Safety (SWS) project, which focuses on aeronautics safety research and development (R&D), plays a crucial role in enabling transformation while ensuring aviation remains the ultrasafe mode of transportation it is today.

SWS’s mission is to explore, understand, and overcome technical challenges to ensure future aviation operational safety. The project accomplishes this through dedicated efforts to determine current and future safety needs, including those identified by pilots, such as new, improved safety solutions and the dissemination of safety knowledge and technology. As the NAS evolves—driven by emerging entrants, markets, technology, and operational concepts—increased traffic volume and airspace complexity may also increase risk. To address this concern, SWS R&D efforts center around both operational and design safety. This article focuses on operational safety and NASA’s role in helping to make certain that, every time you fly, you’re armed with expanded safety awareness and assurance that technical challenges have been addressed.

Operational safety is the systemwide prevention of accidents and incidents through monitoring, assessing, and mitigating risks that may be present during the delivery of a service or conduct of an activity (e.g., potential hazards during aircraft, airport, and air traffic control operations), and it relies on essential safety management systems (SMSs). Airline pilots play a critical role within an airline’s SMS by providing key operational context behind safety events. This operational insight is critical to current and future versions of SMS. NASA’s operational safety R&D focuses significantly on next-generation SMS.

In 2018, the National Academies published a study concluding that today’s SMS must evolve to effectively address aviation safety concerns in the future NAS. The report recommended that NASA, through the SWS project, develop concepts of operation and lead R&D efforts on the design, implementation, and practice of the envisioned future SMS. Working with ALPA and other industry partners, SWS is defining the In-Time Aviation Safety Management System and developing the requirements to address challenges and opportunities in artificial intelligence (AI)–based technologies, including machine learning (ML) and the use of safety-critical data.

The collaborations address barriers to systemwide access, data governance, data sharing, and protection of sensitive safety data. They also involve R&D, testing, and evaluation of NASA-developed AI-based data analytics techniques designed to more effectively utilize systemwide big data. The goal is to anticipate, detect, and assess complex patterns that signal early signs of anomalies and latent threats before they pose risks to system safety. This will result in better products, procedures, and enhanced awareness in the system for pilots and other end users. Importantly, these efforts are guided by human-centered design approaches to develop trustworthy AI-based decision tools and aids.

SWS R&D efforts take many forms, all underpinned by collaboration. Academic and industry partnerships, such as our working relationship with ALPA, are crucial for developing new ways of detecting, analyzing, and acting on safety data. Multiple partner projects are under way, including one to examine what goes right during line operations to better understand what goes wrong. This collaboration has led SWS to identify new safety-producing data types that can serve as data adjuncts to more traditional SMS programs and yield better safety insights. Additionally, these R&D collaborations support new ways of identifying airline training needs and improving instructional methods. For example, one airline has recently licensed our patented flight simulator technologies to enhance flight crew attention and performance improvements.

The SWS team is engaged in significant R&D to develop AI-based data analytical methods that better analyze large datasets for precursors and indicators of potential safety risks. These technologies have the potential to revolutionize aviation safety analysis and, consequently, provide substantial improvements to the safety risk assessment process that airlines use in their daily SMSs. In partnership with an airline that has ALPA pilots, SWS-developed algorithms are employed to analyze emerging indicators of causal and contributing factors behind unstable approaches and the human factors of go-around execution and decision-making.

We’re also developing capabilities that can more effectively analyze text-based safety and nonsafety reports, such as Aviation Safety Action Programs (ASAPs), and have the potential to significantly reduce the time and effort it takes to review such reports. Natural language processing and other AI-based methods can also mine datasets to uncover hidden safety risks that aren’t determinable today due to limited search capability.

The advantages of powerful AI-based technologies extend beyond airlines’ SMSs; they have the potential to serve as a systemwide safety capability. SWS is focusing on the development of capabilities to integrate—or fuse—different types of safety data. Currently, most aviation safety data is held in data silos, such as Flight Operational Quality Assurance, ASAPs, Line Operations Safety Assessments, and other airline SMSs. Significant analytical effort is required to integrate safety findings from unrelated and disparate SMS datasets.

A fused safety dataset can be used to identify the various factors that led to how an incident or accident occurred, how it could have been mitigated or prevented, and if policy or procedures should be revised as a result of these analyses. SWS is helping to develop these capabilities as part of the larger effort to redesign and implement the future Aviation Safety Information and Sharing system (ASIAS 3.0), which is being led by MITRE Corporation.

The forecasted increase in volume, variety, and velocity of aviation safety data and ongoing issues with data veracity (e.g., data noise and glitches) represent significant R&D and system implementation challenges. It will require new ways to ingest, process, manage, and analyze data. In our research partnership with MITRE, the FAA, and an airline with ALPA pilots, we’re helping to define ASIAS 3.0 system architecture and enabling capabilities for identifying future systemic NAS risks.

This R&D effort focuses on integrating our AI-based analytics tools and models while also advancing novel data-fusion methods designed to handle vast, complex datasets. SWS services, functions, and capabilities will support more adaptive analytic strategies and enable automated identification of safety risks, including enhanced predictive modeling of systemwide risks. As a result, the data that pilots and other frontline employees provide will be used to its full potential and allow airlines and regulators to make better decisions based on true operational experiences.

With new data-fused capabilities and advanced big data analytics, ASIAS 3.0 will also enable the automatic fusion of voice and narrative data, thereby vastly enhancing vulnerability detection, supporting safety event analysis, and providing more timely risk-mitigation actions. This partnership will also establish a partitioned data enclave that allows airlines to access ASIAS 3.0 data analytic tools within a data-governed framework. This framework ensures individually secured, proprietary analytical environments in which airlines can access and analyze their own safety data and share with their pilots for increased operational awareness. Through these efforts, the SWS project strives to enable safe adoption of ML capabilities, predictive analysis methods, advanced data mining, and other state-of-the-art technologies.

Advanced aviation safety analysis capabilities are entering a new era of innovation and efficiency. Technological innovation that enables or supports the future air transportation system must be implemented thoughtfully and not outpace the ability to manage and ensure safety. Together with ALPA and our other valued partners, SWS is proving that, when it comes to the NAS, we can innovate for safety and safely innovate.

In an upcoming issue of Air Line Pilot, NASA will discuss how the agency addresses key safety challenges in safety assurance for design.