22 “A controller can handle one or even two uncrewed aircraft in their sector, but if they have several of them and some of the communications latencies and other challenges with an uncrewed aircraft present themselves, it can become difficult. And then, as you go to an m:N operation, it is really hard for a single pilot to have a conversation with multiple controllers at the same time. “And so, we’re going to have to figure out a way to have m:N operations without routine pilot-to-controller voice communication. We have a lot of challenges ahead of us to make that happen.” The experimental m:N UAS operations running today are mainly in the US Federal Aviation Administration’s (FAA) key site around Dallas, Texas, in special airspace, where they don’t have to interact with air traffic control routinely, he notes. The NASA m:N working group brings the UAS community together to discuss technical and operational issues in depth, and the whitepaper is intended as a means of sharing their discussions more widely. Lacher emphasises, however, that the group is not making recommendations for rules or acting as an advisory body, although he anticipates influencing the way standard organisations proceed. The working group is trying to figure out how m:N operations can interact with air traffic control in the future, and Lacher believes human factors are significant here. When a pilot receives an instruction, they don’t just blindly execute it, but evaluate whether they can safely do so, based on their situational knowledge. “It is not a matter of just translating speech to text and then turning that text into a command to be executed by the FMS on the uncrewed aircraft,” says Lacher. “It’s about being able to ensure the safety of flight, and that technology would be significant to invent. So what we have today is a human being listening to these instructions, having the situational awareness to know whether they can be followed, and then issuing commands to the uncrewed aircraft.” Situational awareness For a pilot in an m:N operation to maintain situational awareness for each flight while monitoring multiple communication frequencies is a challenge of a different nature to that faced by an air traffic controller overseeing 15 or more flights in their sector. Lacher argues that an air traffic controller is really managing one thing – a chunk of airspace – even though they have to keep track of many moving parts within it. This doesn’t mean that a pilot managing aircraft in multiple different airspaces could have enough situational awareness for each one to be able to execute instructions safely, all while talking with ATC. Management of multiple UAS by one person clearly needs technological help, possibly in the form of artificial intelligence (AI), although Lacher doesn’t regard AI as a panacea or even as an entity in its own right. “I like to talk about AI as a technology, and about AI-enabled systems,” he explains. “You can build m:N systems without AI, but there are some solutions in which AI makes the most sense for efficient implementation.” Assessing AI One such application that Lacher suggests is runway identification. “AI-enabled machine vision might be required to ensure the aircraft lands appropriately on the runway. It may be a way to augment precision navigation capabilities for landing, as an example, August/September 2024 | Uncrewed Systems Technology In conversation | Andy Lacher Zipline’s UAS-based parcel-delivery service served as a case study for ‘emergent autonomy’ in which groups of aircraft self-organised when coming into land by following simple rules
RkJQdWJsaXNoZXIy MjI2Mzk4