AFMC organizations are working together to increase the safety and efficiency of debris collection at the Holloman High Speed ​​Test Track > Arnold Air Force Base > Article Display

In the New Mexico desert, the men and women of the 846th Test Squadron, 704th Test Group, of the Arnold Engineering Development Complex send test articles attached to rocket sleds that race along rails up to 10 miles long.

Some tests, like ejection seat and target penetration tests, send debris across the desert; cataloged as part of the data collection for the test.

Project Zero is an initiative of the 846 TS and the Air Force Research Laboratory’s Strategic Development Planning and Experimentation Office (SDPE) aimed at reducing the task of identifying and tracking both planned and unplanned debris from tests at the Holloman High Speed to automate Test Track, or HHSTT, with the use of small, unmanned aircraft systems or sUAS, better known as drones, and machine learning algorithms, a subset of artificial intelligence or AI.

“If successful, Project Zero would reduce the time Airmen spend searching and locating debris components and increase safety by ensuring explosives disposal personnel spend less time in an area with dangerous wildlife, unexploded ordnance and desert heat,” said Maj. Ryan Middleton, C-130 navigator and experimenter with SDPE. “Project Zero can also improve data collection and provide a unique perspective for capturing tests.”

The 846 TS used sUAS to get a bird’s-eye view of the track, but it was done in a manual mode, as opposed to the automated, artificially-intelligent operations and data analysis tracked by Project Zero.

“We are always open to automating manual processes and exploring solutions that can increase efficiency and safety,” said 2nd Lt. Aaron Runnells, rocket sled test engineer on the 846 TS. “The idea of ​​using machine learning to train drones to operate in the real world or use them in real-world scenarios can open up new ways of collecting data without the risk of having to be physically on site.”

The HHSTT effort provides an opportunity to compare the use of synthetic data versus the use of real-world data for model training. The team is also investigating how digital environments can be used to understand vulnerabilities and resiliency in machine learning algorithms before deployment.

“The partnership between the Air Force Test Center’s 846th Test Squadron and AFRL’s SDPE Division is a perfect fit; we provide a relevant use case along with decades of data to train models, and SDPE provides the expertise and experimentation team to quickly determine how quickly we can deploy this capability,” said Runnells.

The Project Zero test case scenario involves testing ejection events on a variety of aircraft. If the ejection system is used, the canopy will be fractured, allowing the pilot to be ejected from the aircraft. This creates a debris field that needs to be identified, located and collected.

Project Zero digitally recreates the HHSTT using open source tools from the entertainment and video game industries. Within the digital HHSTT, the ejection system tests are recreated based on physics models and using video game engines and used to train machine learning models to identify, track and report the location of debris.

“The digital environment enables efficient training by modeling different scenarios – environmental conditions, test scenarios, malfunctions – to increase the resilience of the models,” said Middleton. “And thanks to our open-source approach, we’re moving fast — we leverage the global community of software developers for updates, bug fixes, and new tools.”

Conducting research to find scalable ways to ensure artificial intelligence and machine learning models are safe and resilient, like Project Zero, aligns with the Department of Defense’s Responsible AI strategy.

“With its unique mission, multidisciplinary team and scope, the 846th Test Squadron provides the proving ground for machine learning experiments with real-world use cases,” said Middleton. “SDPE is excited to partner with the 846th Test Squadron not only because of the opportunity for experimentation, but also because Project Zero can improve data collection and improve security for the 846th Test Squadron. The open-source digital tools that create the environment can be used by other organizations in the DOD to create their own ML training events [machine-learning] models.”

Runnells agreed.

“If we get this right, we will demonstrate the ability to train a drone with machine learning in a digital environment where simulations can be repeated multiple times, saving time and money,” he said.

Middleton also pointed out how efforts like Project Zero offer opportunities for the DOD to gain support from industries not traditionally seen as part of the defense industry grassroots.

“This is an example of attracting new and non-traditional talent to support AEDC mission sets,” he said. “The innovative vision of the 846th Test Squadron leadership enables new partnerships with industry, and the technical proficiency of AEDC engineers inspires industrial engineers to support critical DOD mission sets.

“Project Zero includes virtual effects artists from across Hollywood’s entertainment industry, as well as roboticists and machine learning experts – some who traditionally would not consider working with the DOD. They were drawn to this effort given the complex technical challenges of the HHSTT and the idea that the results could increase safety for pilots and explosive ordnance disposal personnel.”

Leave a Reply

Your email address will not be published.