10 septembre 2020 | International, Aérospatial, C4ISR

Pentagon to pit AI against human pilots in live fighter trials

and

WASHINGTON — U.S. Defense Secretary Mark Esper announced Wednesday that the Pentagon intends to conduct live trials pitting tactical aircraft controlled by artificial intelligence against human pilots in 2024.

The announcement comes three weeks after an AI algorithm defeated a human pilot in a simulated dogfight between F-16s, something Esper described as an example of the “tectonic impact of machine learning” for the Defense Department's future.

“The AI agent's resounding victory demonstrated the ability of advanced algorithms to outperform humans in virtual dogfights. These simulations will culminate in a real-world competition involving full-scale tactical aircraft in 2024,” Esper said in prepared remarks delivered to the department's Artificial Intelligence Symposium.

The Aug. 20 test was the finale of the Pentagon research agency's AI air combat competition.

The algorithm, developed by Heron Systems, easily defeated the fighter pilot in all five rounds that capped off a yearlong competition hosted by the Defense Advanced Research Projects Agency. Heron's AI system gained notoriety throughout the competition for its aggressiveness and the accuracy of its shot.

But the system wasn't perfect. Heron often made an error in basic fighter maneuvers by turning away from enemy aircraft to where the AI thought the other aircraft would go. It was then unable to recover throughout the fights.

“There are a lot caveats and disclaimers to add in here,” Col. Dan Javorsek, program manager in DARPA's Strategic Technology Office, said after the test, including that the AI had significant information that might not be available in an actual combat scenario.

Military officials have long eyed the potential for AI to control aircraft, whether as part of a “loyal wingman” setup where a number of systems are controlled by one pilot, or through taking existing systems and making them optionally manned.

https://www.c4isrnet.com/artificial-intelligence/2020/09/09/dod-to-pit-ai-vs-human-pilots-in-live-fighter-trials-by-2024/

Sur le même sujet

  • US Navy should turn to unmanned systems to track and destroy submarines

    14 avril 2020 | International, Naval

    US Navy should turn to unmanned systems to track and destroy submarines

    By: Bryan Clark Anti-submarine warfare, or ASW, is one of a navy's most difficult missions. Sonars detect submarines with only a fraction of the range and precision possible using radars or visual sensors against ships above the water. Submarines can carry missiles able to hit targets hundreds of miles away, requiring searches to cover potentially vast areas. And the torpedoes that aircraft and surface ships use to sink submarines need to be dropped right on the submarine to have any chance of sinking it. These challenges led the Cold War-era U.S. Navy to rely on a sequential approach for tracking enemy submarines. Electronic or visual intelligence sources would report when an opposing sub was leaving port, and it would hopefully get picked up by sound surveillance, or SOSUS — sonar arrays on the sea floor — as it entered chokepoints, like that between Iceland and the United Kingdom. Patrol aircraft would then attempt to track the submarine using sonar-equipped buoys, or sonobuoys, and eventually turn it over to a U.S. nuclear attack submarine, or SSN, for long-term trail. The U.S. ASW model broke down, however, in the decades following the Cold War as U.S. submarine and patrol aircraft fleets shrank, the Chinese submarine fleet grew, and Russian submarines became quieter. Today, the U.S. Navy devotes enormous effort to tracking each modern Russian submarine in the western Atlantic. During the 2000s, the strategy of full-spectrum ASW started an essential shift in goals, from being able to sink submarines when needed to being able to defeat submarines by preventing them from accomplishing their mission. Full-spectrum ASW and other current concepts, however, still rely on aircraft, ships and submarines for sensing, tracking and attacking enemy submarines to bottle them up near their own coasts or sink them in the open ocean. Although SOSUS has improved since the Cold War and is joined by a family of new deployable seabed arrays, the next link in the U.S. ASW chain is still a P-8A Poseidon patrol aircraft, an Arleigh Burke-class guided-missile destroyer, or a U.S. SSN. These platforms are in short supply around the world, cost hundreds of millions to billions of dollars to buy and cost hundreds of thousands of dollars a day to operate. With defense budgets flattening and likely to decrease in a post-COVID-19 environment, the U.S. Navy cannot afford to continue playing “little kid soccer” in ASW, with multiple aircraft or ships converging to track and destroy submarines before they can get within missile range of targets like aircraft carriers or bases ashore. The Navy should instead increase the use of unmanned systems in ASW across the board, which cost a fraction to buy and operate compared to their manned counterparts. Unmanned aircraft could deploy sonobuoys or stationary sonar arrays, and unmanned undersea or surface vehicles could tow passive sonar arrays. Unmanned surface vehicles could also deploy low-frequency active sonars like those carried by U.S. undersea surveillance ships that can detect or drive off submarines from dozens of miles away. Although autonomous platforms will not have the onboard operators of a destroyer or patrol aircraft, improved processing is enabling small autonomous sensors to rapidly identify contacts of interest. Line-of-sight or satellite communications can connect unmanned vehicles and sensors with operators ashore or on manned ASW platforms. A significant shortfall of today's ASW concepts is “closing the kill chain” by attacking enemy submarines. Air- or surface-launched weapons have short ranges and small warheads that reduce their ability to sink a submarine, but their cost and size prevents them from being purchased and fielded in large numbers. Unmanned systems could address this shortfall in concert with a new approach to ASW that suppresses enemy submarines rather than destroying them. During World War II and the Cold War, allied navies largely kept submarines at bay through aggressive tracking and harassing attacks, or by forcing opposing SSNs to protect ballistic missile submarines. The modern version of submarine suppression would include overt sensing operations combined with frequent torpedo or depth-bomb attacks. Although unmanned vehicles frequently launch lethal weapons today under human supervision, the small weapons that would be most useful for submarine suppression could be carried in operationally relevant numbers by medium-altitude, long-endurance UAVs or medium unmanned surface vessels. Moreover, the large number and long endurance of unmanned vehicles would enable the tracking and suppressing of many submarines over a wide area at lower risk than using patrol aircraft or destroyers. Today the U.S. Navy uses unmanned systems in ASW primarily to detect submarines. To affordably conduct peacetime surveillance and effectively defeat submarines in wartime, the Navy should increase the role of unmanned systems. Using manned platforms to conduct command and control, and unmanned vehicles to track, deter and engage submarines, could significantly reduce the costs of ASW operations and enable the Navy to scale its ASW efforts to match the growing threat posed by submarine fleets. Bryan Clark is a senior fellow at the Hudson Institute. He is an expert in naval operations, electronic warfare, autonomous systems, military competitions and war gaming. https://www.defensenews.com/opinion/commentary/2020/04/13/us-navy-should-turn-to-unmanned-systems-to-track-and-destroy-submarines/

  • CENTCOM commander eyes future of Middle East | Military Times Reports

    21 juin 2021 | International, Aérospatial, Naval, Terrestre, C4ISR, Sécurité

    CENTCOM commander eyes future of Middle East | Military Times Reports

    Marine Gen. Kenneth McKenzie, commander of CENTCOM, talks about what's ahead for his area of operations now that the Pentagon's eye has turned toward Russia and China. Tension with Iran and the effort to aid Afghan interpreters are also covered in a conversation with Military Times managing editor Howard Altman.

  • Navy Orders Two Hellfire SSMM Systems for LCS

    24 novembre 2020 | International, Naval

    Navy Orders Two Hellfire SSMM Systems for LCS

    Posted on November 23, 2020 by Richard R. Burgess, Senior Editor ARLINGTON, Va. — The U.S. Navy has ordered two more Surface-to-Surface Missile Modules (SSMMs) for integration into the Surface Warfare Mission Package of the littoral combat ships, bring to four the number of SSMMs on order. The Naval Sea Systems Command awarded Northrop Grumman Systems Corp. a 10.7 million firm-fixed-price contract modification for two additional SSMM systems for delivery by November 2022, a Nov. 20 Defense Department contract announcement said. The SSMM is a modular weapons system that fires Lockheed Martin-built AGM-114L Longbow Hellfire missiles from launchers built by Teledyne Brown Engineering. Each launcher houses a total of 24 missiles. A prototype launcher has demonstrated the capability to defend against multiple swarming Fast Attack Craft/Fast Inshore Attack Craft (FAC/FIAC). During one test, a volley of six missiles were launched in rapid succession from an LCS, successfully destroying a swarm of six high-speed targets. More than 100 missiles have been fired to date with a greater than 90% successful engagement rate. The SSMM achieved Initial Operational Capability on the Freedom-variant LCS in February 2019 and was deployed on USS Detroit in November 2019. Northrop Grumman is under contract to build four SSMM systems so far. The Navy has a requirement for 12 SSMM systems. https://seapowermagazine.org/navy-orders-two-hellfire-ssmm-systems-for-lcs/

Toutes les nouvelles