Dr. John K. Lauber
Co‐Chair, Committee on Autonomy Research for Civil Aviation
Aeronautics and Space Engineering Board
Division on Engineering and Physical Sciences
National Research Council
The National Academies
Committee on Science, Space, and Technology
U.S. House of Representatives
January 21, 2015
Chairman Smith, Ranking Member Johnson, and members of the committee:
Thank you for the opportunity to appear before you today in my capacity as the Co‐Chair the National Research Council’s Committee on Autonomy Research for Civil Aviation. Together with my Co‐Chair, John‐Paul Clarke of the Georgia Institute of Technology, I had the pleasure of working with a distinguished group of scientists and engineers from a variety of disciplines and academic, industrial and government settings to develop a national research agenda to support the introduction of what we termed increasingly autonomous elements into our civil aviation system. This study was conducted at the request of NASA’s Aeronautics Research Mission Directorate, and it was conducted over the course of about 18 months; our final report was issued last summer. Whereas our study was requested by NASA, we were asked not to provide recommendations to that one agency but moreover to identify and prioritize elements of a national agenda of research and development that could be pursued by government, academia, and industry. You have copies of the Summary of our report before you, and I will try to succinctly summarize our findings and recommendations below.
Before we get into the details of our findings, I want to summarize briefly the approach we took to this task. Because of our specific focus on civil aviation applications of increasingly autonomous systems, it was necessary to explicitly recognize a couple key characteristics of civil aviation that would set the context for our findings. First and foremost, one hallmark of our nation’s civil aviation system is safety, especially in civil air transport operations. For a variety of reasons many decades in the making, our air transportation system operates at unprecedented levels of safety, and it is clear that the introduction of increasingly autonomous capabilities into that system will be acceptable only if they preserve or enhance this high level of safety and reliability. Secondly, we had to recognize the diversity of aircraft, ground systems, and personnel that comprise our civil aviation system. Because so‐called “legacy” aircraft and support systems will continue to operate for the foreseeable future, it is clear that civil airspace must safely and efficiently accommodate everything from Piper J‐3 Cubs designed in the 1930s to increasingly autonomous, unmanned rotary and fixed‐wing vehicles whose designs and applications are continually evolving. These features and characteristics of civil aviation in the United States were key drivers of the elements of the national research agenda we recommended.
It is also important to understand how autonomous capabilities will be integrated over time into the National Airspace System (NAS). Our committee adopted the view that the civil aviation system of the future will evolve over time starting with a baseline defined by today’s system, particularly with regard to levels of automation that are designed into present‐day cockpits and air traffic management systems. Autonomy, in this context, is a characteristic or feature of future aviation automation systems that we use to refer to aviation systems that will be capable of operations over extended distances and for long periods of time without direct human supervision or intervention. As we point out in our report, this has some profound implications for urgent research in machine vision, perception, and cognition that must be developed to provide the functional equivalent of a “see‐and‐avoid” capability, which is the cornerstone for collision avoidance in our national aviation system. This is one key example of what we mean when we talk of increasingly autonomous systems—systems that will evolve to perform more and more of the functions presently provided by human pilots, controllers, and other skilled aviation personnel.
We started our task by identifying barriers to the increased use of autonomy in civil aviation systems and aircraft. Some of these barriers are technical, some are related to certification and regulation, and some are related to legal and social concerns. Our research agenda was developed to address these barriers.
The Technology Barriers we identified are as follows: (1) Communications and data acquisition requirements in an increasingly autonomous civil aviation system may push the boundaries of bandwidth and spectrum management necessary to support these operations; (2) Cyberphysical security, a topic of increasing concern generally, may be particularly critical to ensure the stability, reliability and functionality of the increasingly‐autonomous civil aviation system of the future; (3) Decision‐making by adaptive, non‐deterministic systems is a critical element of autonomy in civil aviation systems, and there are significant challenges to the design, implementation and testing of such systems at present; (4) As I’ve previously mentioned, the diversity of vehicles and systems that must be accommodated in a civil aviation system will make it more difficult to incorporate increasingly autonomous systems; (5) advances in human‐machine integration are needed because increasingly autonomous systems will require humans and machines to work together in new and different ways that have not yet been identified. (6) I’ve also mentioned machine sensing, perception, and cognition as a significant technological hurdle that must be addressed; (7) Increasingly autonomous systems will present new, and not well‐understood challenges in terms of system complexity and the ability of the civil aviation system as a whole to resist precipitous declines in performance because of isolated failures in one part of the system; and finally, (8) Existing approaches to the formal verification and validation of systems are not adequate to address these requirements in increasingly autonomous systems. Each of these technical barriers is more fully‐developed in the text of our report.
Four barriers in our report are related to Regulation and Certification: (1) Airspace access for unmanned aircraft is a significant barrier for present operations of unmanned aircraft; (2) Today’s certification process doesn’t adequately take into account the special characteristics of increasingly autonomous systems; (3) Many of the safety standards and requirements that are applied to increasingly autonomous systems were developed in the context of crewed, passenger‐carrying aircraft, and it’s not clear how well‐suited they are to assure an equivalent level of safety for unmanned aircraft operations; and (4) Even if we had adequate processes and procedures for verification, validation, and certification of increasingly autonomous systems, the absence of trust in such systems will impose significant barriers to their widespread adoption and utilization.
Finally, our committee identified two other barriers that could seriously impede the degree and speed of adoption of increasingly autonomous technology in civil aviation: These are (1) legal issues associated with public policy, law. and regulation, and (2) social issues, especially public concerns about privacy and safety.
From a consideration of these technical, regulatory/certification, and social/legal barriers, our committee identified eight high‐level research projects that collectively will enable the realization of our vision of increasingly autonomous civil aviation systems and operations. Our report discusses each of these in some depth, but in the interest of time this afternoon, I’ll simply summarize and highlight our recommendations for a national research agenda. I would also point out that the committee feels that each of these research issues is important, but we consider the first four of these to be most urgent and most difficult.
1. Behavior of Adaptive/Nondeterministic Systems: Autonomous systems are characterized by their ability to learn from experience and to adapt to changing conditions. This means that the outputs of such systems can change over time, and that their response to a given set of conditions might be different over time as the systems gain experience. This poses significant technical challenges for design, testing and certification of these systems.
2. Operation Without Continuous Human Oversight: Another defining characteristic of increasingly autonomous systems is their ability to operate for extended periods of time without direct human oversight, that is, without the need for a human to monitor, supervise, and/or directly intervene in the operation of these systems in real time. This will require that the functions currently provided by human operators are accomplished by the increasingly autonomous systems during periods when the system operates unattended. Increasingly autonomous systems will need to respond safely to degradation or failure of aircraft systems as well as other high‐risk situations encountered during a mission.
3. Modeling and Simulation: The committee recommends a significant undertaking to develop the theoretical basis and methodologies for using modeling and simulation to accelerate the development and maturation of increasingly autonomous systems and aircraft. Potential applications include component design, training and coaching of human operators, creating and enhancing human trust in increasingly autonomous systems, accident and incident investigation, and furthering our understanding of cybersecurity vulnerabilities and how to mitigate those risks.
4. Verification, Validation, and Certification: I’ve previously stated our committee’s concerns about the inadequacy of present approaches to verification, validation, and certification of increasingly autonomous systems. It is important to recognize, however, that one of the key reasons for the previously‐noted high levels of safety we experience in civil aviation is because of the formal requirements imposed by the FAA for verification, validation, and/or certification of hardware, software and people, as appropriate.
5. Nontraditional Methodologies and Technologies: The active and growing community of hobbyists and prospective entrepreneurs developing increasingly autonomous unmanned aircraft and associated systems are relying heavily on open‐source hardware and software, resulting in a proliferation of low‐cost, highly‐capable technology. We believe that such technologies will be a key element of the increasingly autonomous civil aviation system of the future, and urge research and development that will allow these technologies to be used safely and efficiently. This would include, in our view, the research that would allow the use of open‐source intelligent software in safety‐critical applications, including unlimited flight operations.
6. Roles of Personnel and Systems: There may be a tendency to believe that the advent of increasingly autonomous systems will lessen the need to assure proper consideration of human factors in the engineering of such systems. However, our committee believes quite the opposite. Because intelligent, adaptive/non‐deterministic systems will require humans and machines to work together in previously unanticipated ways, it is imperative that research be undertaken to further understand the roles and responsibilities of humans and machines, and to address the question of how to safely and efficiently integrate these in an operational environment.
7. Safety and Efficiency: The committee believes that increasingly autonomous systems could enhance the safety and efficiency of civil aviation. Our report discusses a wide range of potential applications of increasingly autonomous technology that could greatly reduce the risk to humans involved in operations such as aerial fire‐fighting operations. We also recognize that such technology might readily be applied to improve the safety of general aviation, especially with respect to single‐pilot operations by, in effect, incorporating an “electronic co‐pilot” that could provide needed, timely assistance to a human pilot. We recommend that this research project include an analysis of accidents and incidents to determine those instances where increasingly autonomous systems might have made a positive outcome possible. Such systems would not be susceptible to factors that adversely affect human performance, such as stress or fatigue. Effective development of increasingly autonomous systems could have a major impact on the need for highly‐skilled, highly‐trained people.
8. Stakeholder Trust: Increasingly autonomous systems can fundamentally change the relationship between people and technology, and one important dimension of that relationship is trust. Even if the necessary developments for verification, validation, and certification are successfully accomplished, in the absence of trust, the potential benefits of increasingly autonomous systems cannot be realized. It is therefore necessary to understand what attributes of systems affect their trustworthiness and how this is communicated to the people who are responsible for their operation.
These then are the major recommendations for research developed by our committee. Although this study was done at the request of NASA’s Aeronautics Research Mission Directorate, we were specifically directed to develop a national research agenda rather than a NASA research agenda. We thus recognize that the research we’ve recommended can and should be addressed by multiple organizations in the federal government, industry and academia. Clearly the FAA has a major role to play, particularly for those elements pertaining to verification, validation and certification. The Department of Defense, although primarily concerned with military applications of this technology, also has a requirement for at least some of their unmanned aircraft to be able to fit seamlessly into operations in the National Aviation System. Each of the high‐priority research projects overlaps to some extent with one or more of the other projects, and each would be best addressed by multiple organizations working in concert. There is already some movement in that direction, however, we believe there is an ongoing need for active coordination of the research effort related to autonomy in civil aviation.
Civil aviation in the United States and elsewhere in the world is on the threshold of profound changes in the way it operates because of the rapid evolution of increasingly autonomous systems. Advanced systems will, among other things, be able to operate without direct human supervision or control for extended periods of time and over long distances. As happens with any other rapidly evolving technology, early adapters sometimes get caught up in the excitement of the moment, producing a form of intellectual hyperinflation that greatly exaggerates the promise of things to come and greatly underestimates costs in terms of money, time, and—in many cases—unintended consequences or complications. While there is little doubt that over the long run the potential benefits of increasingly autonomous systems in civil aviation will indeed be great, there should be equally little doubt that getting there, while maintaining or improving the safety and efficiency of U.S. civil aviation, will be no easy matter. Furthermore, given that the potential benefits of advanced systems—as well as the unintended consequences—will inevitably accrue to some stakeholders much more than others, the enthusiasm of the latter for fielding increasingly autonomous systems could be limited. In any case, overcoming the barriers identified in this report by pursuing the research agenda proposed by the committee is a vital next step. Even so, more work beyond the issues identified here will certainly be needed as the nation ventures into this new era of flight.
John K. Lauber is an independent consultant. He served as Senior Vice President and Chief Product Safety Officer for Airbus SAS in Toulouse, France until his retirement in January, 2008. Prior to assuming this position in January 2005, Dr. Lauber was Vice President – Safety and Technical Affairs for Airbus North America in Washington, DC. From 1997 to 2000 he was Vice President–Training and Human Factors for Airbus Service Company, and prior to joining Airbus was Vice President - Corporate Safety and Compliance at Delta Air Lines. In 1985, Dr. Lauber was nominated by President Reagan and confirmed by the US Senate for a term as a Member of the National Transportation Safety Board. In 1990 he was re-nominated by President Bush and confirmed for a second term at the NTSB, which he served through 1994. Dr. Lauber has also served as Chief of the Aeronautical Human Factors Research Office for NASA Ames Research Center at Moffett Field, CA, where he was instrumental in the development of advanced flight crew training concepts such as Crew Resource Management (CRM) and Line Oriented Flight Training (LOFT) that are now used by airlines around the world. From 1969 until 1972, he was a Research Psychologist at the US Naval Training Devices Center in Orlando, FL. Dr. Lauber holds a Ph.D. degree in neuropsychology from Ohio State University (1969). He is a commercial pilot, with both airplane and helicopter ratings, and is type-rated in the B727 and the A320. He has received numerous awards, including NASA’s Outstanding Leadership Award, the Flight Safety Foundation/Aviation Week and Space Technology Distinguished Service Award, the Industry/Public Service Award from Air Transport World, the Boeing/Flight Safety Award for Lifetime Achievement in Aviation Safety, and most recently, the Joseph T Nall award from the International Aviation and Transportation Safety Bar Association. He has served as President of the International Federation of Airworthiness (2004-2005) and the Association for Aviation Psychology (1984). He has also served on several NASA, FAA and NRC Boards and committees, including the Workshop on Assessing the Research and Development Plan for the Next Generation Air Transportation System, and the Committee on the Effects of Commuting on Pilot Fatigue; presently he is serving as Co-chair of the NRC committee on Autonomy Research for Civil Aviation. He is a member of the MITRE Aviation Advisory Committee and is the Vice Chairman of the Puget Sound Harbor Safety Committee (Public-at-Large member), and was recently appointed to the Safety and Reliability Review Board for Carnival Cruise Lines.
An archived webcast of the hearing can be found on the House Science, Space, and Technology Committee's Web site.