The data-driven motor pool

U.S. Army tests AI & analytics that predict when vehicles need fixing

Bradleys unloaded at Busan, Korea

The U.S. Army unloads Bradley Fighting Vehicles in Busan, Republic of Korea, as part of a prepositioning initiative on the peninsula. (U.S. Army photo by Spc. Bryan Willis)

What if your car could tell you exactly when it needs repairs?

Using artificial intelligence and machine learning, your car’s computer could one day analyze your driving habits and predict when you'll have to bring it in for service. It would also factor in variables like engine age, weight, road conditions, oil viscosity, idle time and mileage, then use advanced algorithms to tell you when it will be time for a tuneup or to change the timing belt.

What works for your car could also work for a fighting vehicle. The U.S. Army is investigating predictive and prognostic maintenance for its fleet of vehicles, using AI and machine learning to flag failing parts and systems before they break down.

“Today, if something breaks, we fix it,” said Kurt Stein, a Raytheon Technologies senior logistics adviser and a retired U.S. Army major general. “That’s the way we’ve done things in the Army for a long time – reactive maintenance. What [we're] trying to do with their new system is anticipate failures before they become problems.”

Raytheon Intelligence & Space, a Raytheon Technologies business, which has expertise in data analytics and maintaining military equipment, is demonstrating data-driven technology to keep the Army's fleet up and running with minimal downtime. The company is collaborating with several commercial companies to upload diagnostic data from sensors already aboard the Army's Bradley Fighting Vehicle and M88 Recovery Vehicle, and forecast equipment failures before they occur.

“The last thing that I’d want to do as an Army leader is put soldiers in a Bradley, send them into harm’s way and have the dang thing break down,” Stein said. “...when in fact we could’ve put a box on it with the latest and greatest of technology, connected to it, and it could’ve told me this or that component is going to become non-operational in the future.”

If the data-driven system signals that a part may become faulty, commanders can repair or replace it before the entire vehicle is compromised. It also gives commanders a better sense of the operational readiness of their vehicles.

“Let’s say a commander has a dozen Bradleys on hand but needs three of them for a specific mission,” said Kevin Frazier, Raytheon Intelligence & Space Sea Logistics and Support business development manager. “How do you choose them? Today, it’s intuition…a guess. With real-time data, commanders could choose the vehicles that are the most operational and mission-ready.”

A data-driven system could also help reduce the supply chain and logistical footprint at forward-deployed locations, or the “edge,” as it’s called.

“Today, the Army is using historical data to decide what parts they should bring with them,” said Butch Kievenaar, RI&S Missions Systems and Solutions business development director. “With AI, machine learning and analytics, Army maintainers can predict what they’ll really need. Instead of dragging along 120 parts, they might need to take only 80 parts. And these will be parts they need, not just good to have.”

The technology could also save the Army time and money by avoiding unnecessary preventive maintenance. Every Bradley gets a weekly maintenance check plus quarterly, semi-annual and annual inspections.

“We know that different parts fail at different rates depending on the environment and the type of operation it’s conducting,” Kievenaar said. “If you’re in the desert, we know sand takes its toll on components, and if you’re doing gunnery operations, your fire control systems will fail faster. Now, we’ll have data to make those informed decisions on when and what needs to be fixed.”

While cost savings, efficiency and a smaller logistical footprint are all benefits of the new technology, mission readiness and soldier safety are the top priorities.

“If you know a component could potentially break based on history, analytics, data and fact, then do you want to put a solider in harm’s way in that vehicle or do you want to use a different vehicle or do you want to swap out that component?” Stein said. “In some cases, you have to go with what you’ve got, but I think this capability provides leaders with better information so they can make an informed decision on whether to send soldiers into combat.”

Published On: 03/12/2020