There is much interest in autonomous systems and their operations in the national airspace system. When it comes to autonomous systems, it is critical to understand various levels of proposed autonomy, reasons for selecting autonomous capabilities, certification approaches, and various challenges (or research needs) associated with the integration of autonomous systems into the airspace system. The initial article aims to cover broad overview rather than any specific element. It must be noted that there are various levels of autonomous systems proposed for different types of aircraft. These include single pilot operations, remotely piloted operations, fully autonomous operations, and many aircraft managed by a few operator pilots.
Reasons for Proposing Autonomous Systems
Long aircraft stage length: There are several reasons for proposing autonomous capabilities. For long-range cargo aircraft, the fatigue rules require pilot changeover after their shift. While the stage length of aircraft is increasing, as high as 17 hours, such flight requires at least two sets of crews each way. An immediate round trip requires additional crew. Such requirements make it harder to schedule entire operations, adding complications. The question is whether we could safely use a single pilot during long oceanic cruise portions of a flight. Such possibilities may reduce the number of crew required for a long-duration, long-haul trip.
Mid-size cargo aircraft: Several start-up companies and cargo operators are experimenting with the ability to serve regions through increasing autonomous flights. Their path typically starts with an on-board safety pilot, then migrating to the remotely operated vehicle, eventually moving towards more aircraft to one pilot ratio. The research, development, and testing activities are underway to set requirements that could be accepted by regulatory bodies across the globe to safely enable such increasing autonomy.
Advanced Air Mobility/Urban Air Mobility: It is argued that many advanced air mobility (in particular, urban air mobility) operations need to be affordable to masses. It is argued that the economics of operation may require autonomous operations. The question remains: how can we scale these operations in a safe manner with a combination of autonomy and humans?
Small drones: For obvious practical reasons, there is no on-board pilot, which requires that the small drone be managed remotely, fully autonomously, and eventually with more drones managed by a limited number of pilots, where the ratio of drones to pilots is more than one. Research is underway to understand the roles, responsibilities, functional and technology requirements, as well as human-autonomy collaboration. Small drones may operate under the unmanned aircraft system traffic management by digital exchanges of intent information and under share and care environment.
Pilot Shortage: As the aviation continues to grow, it is expected that the number of pilots will be inadequate to meet the expected demand. The question remains: how can we safely grow aviation industry if we continue to have pilot shortages?
Airspace Integration Considerations
Pilots have been instrumental in the safety of operations. We have seen several examples and heard many interesting stories how pilots saved the flight in emergency situations.
In order to accept increasingly autonomous flight and operations, we need to also expect similar levels of safety support during flight, particularly for off-nominal conditions. The particular challenges for airspace integration are as follows:
- Air Traffic Controller Workload: Fundamentally, growth could be enabled in aviation with various levels of autonomous aircraft operations if air traffic control system and air traffic controller workload is not unsafely increased. It is possible that an increased level of decision support tools and automation could also be made available to enable such operations. However, similar to autonomous systems of aircraft, research is needed to ensure safe and efficient introduction of automation for air traffic management.
- Communications and Delay: The question is – will an air traffic controller notice a difference in communication durations while communicating with a remote pilot, single pilot, or a fully autonomous flight whose systems may require satellite communications? If so, would that matter (and how) in the way airspace operations are managed by the air traffic controller – particularly under off-nominal conditions (e.g., large-scale weather disruptions), emergencies or contingencies (e.g., bird strike).
- Off-nominal, emergency, and contingency operations: How would an emergency for an autonomous aircraft be managed where an air traffic controller will have to support the affected aircraft operation as well as other surrounding aircraft by clearing the way for affected aircraft?
- Reliable Voice-to-Text: If a fully autonomous aircraft or drone is expected to operate without any supervising pilot, voice-to-text technology must reach high maturity for the aviation environment. To date, perfect voice-to-text for aviation remains a research area.
- Mix equipage operation: At one time, I used a photo from a suburban road in India, where I grew up. The picture showed a bus, a rickshaw, a bicycle, a moped, a motorcycle, a car, pedestrians, and a cow. The entire road would then operate at the speed of that cow, whose intent is hard to fathom. Similarly, mixed equipage remains an interesting consideration. The airspace system could be made safe and efficient if the intent is available from all parties that operate in the airspace. Furthermore, speed profiles, climb and descent rates, turning radii of increasingly different mix of aircraft needs to be well represented in the air traffic management system to ensure predictability of conflicts, understanding of choke points, and support they may need from air traffic control.
- Uncooperative aircraft: One of the key challenges for integrating fully autonomous or remoted piloted aircraft in the airspace system is ability to stay clear of visual flight rules following general aviation (particularly aircraft with no radio and no beacon) and military aircraft as those are not managed by air traffic control system. A conflict detection and resolution capability to stay clear of non-cooperative aircraft is critical. Research is needed to finalize requirements and means of compliance for such capability.
- Expected level of services by increasingly autonomous aircraft: Collectively we need to understand and define the level of services expected by various levels of increasingly autonomous aircraft from the air traffic control system. These possible services include, but are not limited to: demand/capacity imbalance and flow management for airspace, airport, or vertiport constraints; strategic deconfliction so that overreliance on tactical deconfliction is reduced to manage workload of any human (remote pilot or controller) that may be responsible for such activity and efficiency of flight is maintained; tactical deconfliction and responsible party; emergency management and coordination with other air traffic; large-scale disturbance management; and collision avoidance. Depending on the level of autonomy and size, weight, and available power of the aircraft, the support needed from the air traffic management system may vary. However, research needs to finalize the requirements and performance standards for on-board equipage and level of air traffic management support that defines air-ground integrated system characteristics. Additionally, understanding who’s flying in the airspace is key to operators. Currently, information about general aviation and military flights is not directly available to operators. Without a full picture of the airspace, the autonomous aircraft will be unable to safely navigate, or may need additional services from air traffic management.
Unmanned Aircraft System Traffic Management: Enabling small drones (e.g., below 55 lbs. and below 400 ft):
The Unmanned Aircraft System Traffic Management (UTM) environment could support small drones and UAM/AAM operations without overloading air traffic control operations. UTM trials demonstrated that it is possible to scale the operations without overloading the air traffic control by using cooperative, digital, intent-sharing, service-oriented architecture with possible roles for third-parties and using a management-by-exception paradigm. However, it remains to be seen if such paradigms can be used in various classes of airspace and a mix of increasingly autonomous and crewed operations.
Today, various classes of airspace experience different mixes of air traffic. There are two possible considerations as we continue to increase the air traffic mix and density. First, aircraft operators want as much flexibility as possible, whereas structure makes it easier to manage the traffic as it increases the predictability of the air traffic management system. The question is whether we can migrate towards flexibility where possible, and structure when necessary. As a result, the type of environment and adjustments to airspace could be bit more dynamic.
The second consideration is integration and interoperability where possible and segregation where absolutely necessary. Integration and interoperability refer to all types of aircraft operating harmoniously in the same airspace at the same type. In some cases, integration and interoperability is not practical. For example, commercial space launches and drones may not operate at the same time close to each other and segregation may be needed. The research question is, how much integration and interoperability is possible with drones, UAM/AAM, and conventional crewed aircraft of various sizes? The basic assumption is segregation reduces the flexibility for many operators whereas integration does not. There is a clear balance between flexibility and safety in some situations as seen in the example of commercial space launches.
In summary, there are various levels of autonomous flight and operations based on their equipage, and functions. Some of these functions are limited by the size of the aircraft or where they are likely to operate along with other vehicles.
Closing Remarks
There are many technology-, operations-, and economics-related reasons to enable increasing autonomous flight and operations of various sizes. The goal of this article is not to promote the view that autonomous flight and operations are ready to take off and should be allowed. Instead, it highlights the complexity of issues that are involved in enabling autonomous operations. While it includes technology, procedures, training, policy, and other job-related considerations, further research on airspace integration is needed to understand how such operations could be accommodated in a widespread, yet safe, manner within the context of other operations that will co-exist.
One thing is very clear: research, development, and testing must conclusively prove the safety of all levels of autonomous flight and airspace operations, particularly under off-nominal conditions.
Parimal Kopardekar, Director, NASA’s Aeronautics Research Institute and Senior Advisor for Airspace Integration, NASA’s Ames Research Center in California’s Silicon Valley.