IBM (Booth #933) is using its Watson and other in-house systems to deliver the Cognitive Equipment Advisor capability to the US Army. The service project, explained to the author at this symposium, complements Army Chief of Staff Gen. Mark Milley’s number-one priority of readiness. While MONCh readers typically read about artificial intelligence (AI) envisioned to enhance training and other aspects of individual and unit readiness, “equipment readiness was also a natural fit for AI,” Charles Waters, an Associate Partner and Account Manager in the Army Logistics & Cloud Services division in IBM’s Global Business Service’s office, said.
IBM is bringing its AI capability to bear, in this instance, to help extend the service life of the Army’s STRYKER fleet. In a recently completed, successful pilot project with the Army’s LOGSA (Logistics Support Activity), IBM took collected data which, heretofore, basically sat in data tables, and examined whether there were performance or activities occurring on the vehicle, to permit the Army to get ahead of problems before they result in materiel failure, or view wider-ranging, systemic problems.
“We’ve used our equipment advisor technology to look at sensor data from the Stryker vehicle platform – we’re talking about 15,000 rows of data that we have gone through in about six weeks, to look at vehicle anomalies and determine where opportunities were, to either intervene before a component breaking, or to look at it and say, instead of an equipment problem, there may be a training or other problem to impact the vehicle and its readiness,” the industry veteran further explained.
The above “structured” data used in the STRYKER pilot project included sensor-furnished content from major onboard components – steering, transmission, engine, fuel and suspension. The IBM pilot project included 10% (about 350) of the service’s nine variants of STRYKER and provided analytics down to the individual vehicle bumper number.
And beyond structured data, IBM’s AI capability further captures “unstructured data” – the notes from a maintainer’s notebook, safety of use messages and other resources, for example - and with good reason.
Mr. Waters added that IBM is able to consume the two data types and join them together to help look through and identify anomalies. He continued: “From an AI and analytics perspective, it is very hard to make a prediction off ‘of a straight line’. You have multiple angles, where unstructured data intersects structured data and provides a total picture. And then to be able to provide the prognostic analytics, we have to look at historical trends and look at the predictive nature of the vehicle or a system. That certainly applies to mission readiness. That’s the capability we bring.”
Asked about other returns on investment from the successful STRYKER pilot project, Mr. Waters noted it permits part maintainers to complete part predictions and ordering ahead of time to maintain a level demand on the supply chain. Beyond that, “we also identified about $(US)11.6 million worth of engine cost avoidance – to repair or replace – in one calendar year.”
IBM’s AI pilot project and its analytics capabilities were also used in an adjacent application – air clearance authority – the process the army determines what mode of transport is used for repair parts. “It took about nine days to do this, but we looked about $26 million worth of cost savings they could have avoided if they had anticipated a part being in one place instead of having to expedite it by way of airlift – it could have perhaps, been transported by ship.”
The next phase of the IBM AI project will soon transition into a production environment, and will be expanded for real-world testing with a STRYKER brigade as it deploys to a training venue.