but I don’t see AI as the only way we’re going to get m:N.” He notes that most of the AI systems intended for use in this area rely on probabilistic algorithms, which means they will not always be perfect. “You have to design your AI such that it not only performs the intended function but is unlikely – within an acceptable level of risk – to have unintended behaviours. And that’s the key.” One established way of achieving this with safety-critical software is by using runtime assurance, which constantly monitors the behaviour of the system to keep it within appropriate bounds. Emergent autonomy However, AI is not required for complex behaviour or autonomy. Lacher recalls writing a paper with a couple of industry colleagues on emergent autonomy, so called because it described a situation in which a group of deterministic systems interacted in a complex manner such that the behaviour that emerged, while safe, could not be predetermined. As a case study, the paper took Zipline’s UAS-based parcel-delivery service with m:N control. “The aircraft would self-organise when coming in to land. No one was telling them what to do, but they created a traffic pattern. There was no AI in that,” he says. “They were all very deterministic capabilities. To a human being watching it looked like autonomy because it was so complex.” Asked what he expects of AI in UAS integration in the foreseeable future, Lacher believes the pace of technological change is likely to accelerate, but he notes the aviation industry’s inherent caution. “Aviation culture is one of determining that something is safe, and safety is a system property. It’s not a property of the technology itself. If you have an AI-enabled system, you have to determine that the whole system is safe, so I think AI is going to emerge more and more as part of systems but in limited implementations. “The first place we are likely to see it is in machine vision used to augment the navigation capabilities of aircraft. We are seeing this in GPS-denied environments in conflicts like Ukraine today and, unfortunately, the systems are hitting their targets in a warzone. The technology is maybe not ready for a safety-critical application, but it is being used, and it’s a very low-cost technology.” As to his remaining ambitions, Lacher wants to continue contributing to important changes in aviation, having been involved in the FAA’s NextGen modernisation of the National Airspace System with collaborative decisionmaking, in which the organisation and airlines shared information to improve air traffic flow, and in the development of the Small UAS Rule while at Mitre. “I hope to be part of another major operational change,” he says. “This notion of a digitally-enabled, cooperative operating mode is perhaps that paradigm. And maybe I’ll be part of that or other mechanisms to change the future.” 23 Uncrewed Systems Technology | August/September 2024 Born in 1964 in New York City, USA, Andy Lacher moved as a small child with his parents to Long Island, growing up in the small town of Valley Stream. This location triggered his interest in aviation as it was under the flight path to JFK Airport. Lacher recalls sitting on the roof of his house with a friend from next door, wielding binoculars and a radio receiver, and listening to pilots and air traffic controllers, and watching the aircraft. “He wanted to be a pilot and I wanted to build airplanes – well, I’m close.” He credits the Valley Stream public schools with giving him a great education, and instilling his enthusiasm for mathematics, science and history, particularly that of the American Civil War. The story of Swedish engineer John Ericsson, who designed the first ironclad warship with a revolving turret, USS Monitor, reinforced his interest in engineering. He also got into photographing wildlife, particularly birds; a hobby he still pursues. Bachelor’s and master’s degrees from George Washington University followed, in electrical engineering and systems engineering, respectively, the latter while working full-time. Lacher began his career at Mitre Corporation in 1987, working on problems such as information sharing in civil aviation. He left in 2001, for jobs with Orbcomm and Intelix, before rejoining Mitre in 2001, where he spent the next 18 years leading roles in UAS and autonomy integration. Roles with Noblis and then Boeing followed, before he joined NASA in October 2021. Lacher has also served on UAS-related panels and committees for organisations such as the FAA and the National Research Council. Andrew R. Lacher
RkJQdWJsaXNoZXIy MjI2Mzk4