U.S. Air Force trial
The U.S. Air Force is testing a combination of drones, artificial intelligence, and cloud collaboration to shorten the time needed to inspect aircraft for wear.
Trial setup and goals
In a trial supported by Boeing and drone operating system developer Near Earth Autonomy, Air Force technicians at Pearl Harbor-Hickam Joint Base, Hawaii, are launching camera-equipped autonomous drones to document the condition of Boeing C-17 transport aircraft. The objective is to reduce the complexity of inspections while improving the accuracy and reliability of the data. Officials say traditional external inspections can take several hours and still miss small details, whereas a drone-centered approach driven by pattern recognition and 3D models is much faster and transmits validated observations to a securely stored cloud.
Operational benefits and ACE concept
“A preflight inspection can currently take up to four hours. We can do it in 30 minutes. That saves pilots a lot of time and ensures aircraft availability,” Near Earth Autonomy manager Alli Locher said on June 27. “Ultimately, wherever you are in the world, you can pull up the tail number, click any location on the aircraft 3D model, and view the image history for that part across the aircraft lifecycle.”
The Air Force is promoting a concept called agile combat employment (ACE) as it prepares for potential operations in the Indo-Pacific and Europe. ACE envisions a hub-and-spoke basing layout: some hubs are larger and fixed, others smaller and mobile.
Analysis, results, and future work
“This approach will further distribute manpower and expertise, resources that are already in high demand, so having a reliable, centralized way to collect and assess aircraft condition will become more important,” said Scott Belanger of Boeing.
“These images are effectively real time, sent to a cloud environment where they are analyzed by Near Earth Autonomy software and our automated damage-detection software,” he added. “We are not trying to replace manual inspections. We are working to enhance human inspection capabilities.” In tests, the drones and associated procedures detected up to 76%–78% of damage, outperforming human detection rates of about 50%.
Looking ahead, Boeing and Near Earth Autonomy are pursuing additional payloads to detect subsurface damage and plan to add more aircraft types to the inspection roster. Locher said Lockheed Martin's C-5 was recently programmed, and Boeing's KC-135 and KC-46 may follow. “Our secret is an autonomous backend on the drone that always knows its position relative to the aircraft, not just the surrounding environment,” she said. “With that, you can operate almost any sensor and obtain a mapped output for that sensor.”
