美U2侦察机首次装备AI副驾驶系统(英文)

2020-12-19  智邦网

编译 致远

据defensenews网报道

名为Artoo的人工智能副驾驶系统于12月15日首次装备洛马公司研制的U-2侦察机,飞行员将与AI副驾驶协同飞行、并肩作战。Artoo将全权接管该机两大关键系统,承担控制侦察机雷达和传感器系统任务。

在加州比尔空军基地飞行训练中,Artoo的任务是找到用于模拟导弹袭击的导弹发射装置,起飞后全权负责传感器的操作和战术导航系统;而U-2侦察机飞行员则专注寻找敌机,并与“AI副驾驶”共享雷达信息。

美国空军采购部门主管威尔·罗珀表示:“Artoo系统也有优缺点,让人类和人工智能为未来算法战争新时代做好准备,是下一步重点工作。”

Artoo系统由U-2侦察机联邦实验室研发,该实验室在10月对U-2侦察机升级中配备了该系统,这在美军尚属首次。

美空军表示,该系统也可修改后用于其他作战飞机。

美空军参谋长查尔斯·布朗上将称:“为在未来与对等对手冲突中作战致胜,必须拥有决定性作战数字优势”“人工智能将发挥关键作用,必须加速变革。”


Artoo, take the wheel: U-2 spy plane flies for the first time with an AI co-pilot

 In Star Wars, Luke Skywalker destroys the Death Star with the adorable, wisecracking droid named R2-D2 in the back of his X-wing, helping navigate and fix the ship in real time.

Now, the U.S. Air Force has its own Artoo — called, well, Artuμ — an artificial intelligence system that flew onboard a Lockheed Martin U-2 for the first time Dec. 15 and was given control of the spy plane’s radar and sensor systems.

On a reconnaissance training mission conducted out of Beale Air Force Base in California, Artuµ was tasked with finding adversarial missile launchers during a simulated missile strike, and it was solely responsible for sensor employment and tactical navigation after takeoff, the Air Force said in a news release.

The human U-2 pilot, referred only by the callsign “Vudu” for security reasons, concentrated on finding enemy aircraft and shared the use of the radar with the AI co-pilot.

“Like any pilot, Artuμ (even the real R2-D2) has strengths and weaknesses,” Air Force acquisition executive Will Roper tweeted in an announcement of the news. “Understanding them to prep both humans and AI for a new era of algorithmic warfare is our next imperative step. We either become sci-fi or become history.”

Artuµ was created by the U-2 Federal Laboratory, which in October successfully updated the plane′s software while it was flying — a first for the U.S. military. The event was made possible by deploying Kubernetes, an open-source, containerized method for automating software updates.

Artuμ is based on a gaming algorithm known as µZero, which has been used to beat human players in chess and Go, Roper explained in an op-ed on Popular Mechanics. The U-2 lab specially trained the AI co-pilot to manipulate the U-2′s sensor suite during “over half a million” computer-simulated missions, according to the Air Force.

“With no pilot override, ARTUµ made final calls on devoting the radar to missile hunting versus self-protection,” Roper wrote.

Although Artuµ was developed to take away from the pilot’s workload in a U-2, it can be modified for use by other combat planes, the service said.