PhysicsAI的动态

"The US Air Force Test Pilot School and the Defense Advanced Research Projects Agency (DARPA) claim to have achieved a breakthrough in machine learning by demonstrating that AI software can fly a modified F-16 fighter jet in a dogfight against human pilots. The claims rest on the USAF and DARPA implementing machine learning in an X-62A VISTA, a plane built as a testbed as it can mimic the performance of other aircraft, and recognition of their work as one of four finalists for the National Aeronautic Association's 2023 Robert J. Collier Trophy, an annual award for exceptional feats of aeronautics or astronautics in America. "The potential for autonomous air-to-air combat has been imaginable for decades, but the reality has remained a distant dream up until now,"?said?Secretary of the Air Force Frank Kendall. "In 2023, the X-62A broke one of the most significant barriers in combat aviation. This is a transformational moment, all made possible by breakthrough accomplishments." DARPA has been testing AI agent software for piloting simulated planes for several years. Its?Air Combat Evolution (ACE)?program dates back to 2020, when?AlphaDogfight trials?pitted human pilots in a flight simulator?against an AI opponent. The AI software won that competition but had an edge – it was allowed to fly at speeds that would have overstressed a real aircraft and generated g-forces that would harm a human pilot. Heuristic or rules-based autonomy has been a common approach in military and space applications. These sorts of expert systems boil down to if-then statements that specify condition-based triggers that lead to specific actions. But this approach is less useful when there are too many variables and rules to account for. "The machine learning approach relies on analyzing historical data to make informed decisions for both present and future situations, often discovering insights that are imperceptible to humans or challenging to express through conventional rule-based languages,"?explained?Daniella Rus, director of MIT CSAIL, in a DARPA video you can see below. "Machine learning is extraordinarily powerful in environments and situations where conditions fluctuate dynamically making it difficult to establish clear and robust rules." https://lnkd.in/guF5DniR

Real Aircraft. Real Autonomy. Real Flights. “We’ve proven a method for going from our laboratories to live airplanes. We’ve proven the method to both get the technology there, but also for our people to support some of the compliance and testing and verification pieces you need to do. We’ve proven that model, that reinforcement learning can fly and it is successful. And we’ve proven that the way we approach the problem is scalable.” … and we’re just getting started. Our aim is to put one million AI pilots in the skies in the next ten years. DefenseScoop's Mikayla Easley spoke to our EVP of Product, Brett Darcey about the SECAF’s recent autonomous flight in the X-62 VISTA; how plannable, directable and trustable autonomy really “works;" and how we’re maturing our AI pilot in anticipation for future programs aimed at bringing autonomy to the Air Force’s manned and unmanned platforms. https://lnkd.in/gSrE9YTU The greatest victory requires no war. #AI #Aviation #GreatestVictory #BuildwithShieldAI #AirForce

The incredible technological advancement demonstrated by the tests conducted on the X-62 VISTA showcases the evolution of AI technology and capabilities in the application of airpower. This is beginning to redefine the rules of the game as we know them. #AI #buildwithShieldAI

We are looking at what’s being claimed as a breakthrough in the use of AI in aerial combat. Quite often, the biggest leaps of advancement happen in extreme testing environments, be it F1, combat aviation or space/ ocean exploration. Such evolutions find excellent use cases in advancing technology that would benefit humanity at large (at some point). What isn’t clear to me here is when they suggest (in the video within the attached link) that the usage restrictions of AI/ ML in air combat was primarily due to the risk of the model’s decision-making process not being ‘understandable or verifiable’. Which they seem to have addressed with guards that limit decisions within the flight safety envelope (at least from what’s disclosed). Aren’t the risks of understandability and verifiability still alive then? Or is free decision-making within a safe envelope the present and future of human-AI interaction? #airisk

?? #RevolutionizingAerospaceThroughAI: Navigating the Future Together ?? Dear fellow aerospace aficionados, As we embark on our quest for innovation in the vast expanse of aerospace, one technology stands out as a guiding star: Artificial Intelligence. With its transformative potential, AI has become the compass guiding our exploration, the wings propelling our aspirations to new heights. In every facet of our industry, from aviation to space exploration, AI emerges as the catalyst for unprecedented advancements. Picture AI-driven autonomous drones mapping uncharted territories, adaptive flight control systems optimizing fuel efficiency, and predictive maintenance algorithms ensuring the safety and reliability of our aerospace fleets. But the marriage of AI and aerospace transcends mere technological synergy; it embodies a profound shift in our approach to exploration and discovery. With AI as our copilot, we navigate the complexities of space missions with greater precision, unlocking mysteries of the cosmos that were once beyond our reach. As we converge on the frontier of AI-driven aerospace, let us embrace the possibilities it offers with open minds and boundless imagination. Together, we chart a course towards a future where AI not only enhances our capabilities but also deepens our understanding of the universe and our place within it. #aerospaceengineering#futureofflight #ExplorationUnleashed

Recent publication on MDPI Sensors ?? A Deep Learning Approach for Surface Crack Classification and Segmentation in Unmanned Aerial Vehicle Assisted Infrastructure Inspections Sensors?2024,?24(6), 1936;?https://lnkd.in/ghytZeSn

Machine Learning in Aerospace - I’m darn proud of our team’s collaboration with MIT Lincoln Laboratory, EpiSci, and AFWERX demonstrating safe and rapid development of machine learning autonomy to the Secretary of the Air Force. This is a senior leader who can speak on AI with first-hand experience. I can’t emphasize enough how important it is that we’re figuring out how to safely and rapidly develop, test, and fly machine learning autonomy. There is no pathway to certify machine learning autonomy for flight-critical systems. This serves as a model to change that. The Defense Advanced Research Projects Agency (DARPA) and United States Air Force Test Pilot School teams are paving a future for aerospace autonomy for defense and civil sectors. #MachineLearning #AI #Autonomy #FutureOfFlight

Leonardo and Daedalean complete flight testing of Artificial Intelligence capabilities for advanced navigation in the rotorcraft domain Leonardo put Daedalean’s cutting-edge, AI-enabled avionics systems to the test to increase safety today and to take a further step toward autonomous flight. Daedalean, a developer of safety-critical and certifiable artificial intelligence systems for situational awareness and flight control, has announced the completion of flight testing of Daedalean’s visual awareness system by Leonardo, a leader in aerospace, defence, and security. "Leonardo is working towards prudently integrating AI in its products and services through both in-house developments and cooperations. By collaborating with emerging companies on predefined use cases, we keep maturing our technology roadmaps towards a safer, affordable and sustainable flight experience," said Mattia Cavanna, Head of Technology & Innovation at Leonardo Helicopters. "Improving situational awareness through system like Daedalean's in the near future could contribute to further prevent aviation accidents and progressively enable higher degrees of autonomy to our platform." Leonardo has shown its commitment to developing and adopting new technologies. The Leonardo Labs serve as technology hubs, connecting emerging talent from leading universities with Leonardo’s veteran experts to drive innovation and identify practical applications. Leonardo Labs pursues research in areas from materials and quantum technologies to sustainability and applied artificial intelligence. Daedalean provides what the company terms Situational Intelligence – the ability to understand and make sense of the current environment and situation, and also anticipate and react to potential threats. Its visual awareness system employs a form of artificial intelligence called machine learning, which leverages recent increases in computer power to do more quickly and effectively what before could only be done by people. “Daedalean is proud to bring our experience creating machine-learned algorithms for aviation to such a prominent player in the world of aviation,” said Daedalean CEO Dr. Luuk van Dijk. “It shows there is growing interest in and understanding of the benefits machine learning can bring today to increase flight safety.” #Leonardo #Daedalean

Unveiling the Future of Defense: Drone GX Prototype Revealed ???? We're thrilled to showcase the first glimpse of our revolutionary Drone GX - a testament to Indian innovation in aerial defense! Drone GX integrates cutting-edge technologies to redefine battlefield awareness: Quantum Well Infrared Photodetector (QWIP): This next-gen sensor boasts unparalleled long-range target detection with exceptional detail, ensuring no threat goes unnoticed. Hierarchical Deep Neural Networking Algorithm: Our proprietary AI empowers real-time threat identification and analysis, enabling precise decision-making in critical moments. Drone GX isn't just a machine - it's a force multiplier for our nation's defense. #DroneGX #DefenseInnovation #MadeInIndia #JaiHind #AerospaceDefense #DefenseInnovation #MilitaryTechnology #FutureofDefense #DefenseIndustry