Cloud, AI drives Army's conditioned-based maintenance
The Army is leveraging cloud computing and artificial intelligence to help commanders quickly identify and anticipate mechanical failures in vehicles, assess equipment functionality and expedite analysis for near real-time decision-making.
For the condition-based maintenance (CBM) initiative, the Army is working with IBM, which is providing the cloud services, software development and cognitive computing aspects for the wireless sensor-based analysis project under the Logistics Support Activity contract.
“We see huge potential benefits with cloud computing from artificial intelligence,” said Col. John Kuenzli, LOGSA commander.
CBM is engineered to transform Army maintenance and transition from a schedule-based "operate-to-failure" system to a "repair-when-needed" approach. AI-enhanced analytics engenders a greater ability to predict anticipated failures.
“CBM examines an entire system from a physics-of-failure perspective joined with data analytics to perform maintenance when there is evidence of need,” Kuenzli said.
The initiative is also designed to give stakeholders the ability to see trends, predict outcomes, evaluate courses of action and adjust tactics when necessary, said Army spokeswoman April Cunningham.
Cunningham explained that the next technical step is to develop "store-and-forward" capability to aggregate platform data until it can be retrieved wirelessly or through a handheld device.
“This capability will enable data collection and analysis in most environments and significantly reduce the bandwidth to transmit vehicle usage data to the enterprise,” Cunningham said.
IBM explained that its Watson computer can instantly access relevant historical data pools needed for data analysis.
“Watson brings an analytical edge to CBM,” said Greg Souchack, manager of service cloud solutions, IBM. “It can read and analyze pools of information to determine when or if something will malfunction or reach the end of its service life. Watson can operate in near real-time and provide continuous updates.”
One analyst explained how data aggregation and analysis, along with greater data sharing, access and interoperability, is highly dependent upon greater migration to cloud technologies.
“The cloud allows you to bring together previously isolated forms and sources of data together, allowing you potentially to draw new insights,” said Peter Singer, strategy expert and senior fellow at The New America Foundation.
The Army recently experimented with these emerging technological methods using cloud computing, data analysis and AI to track CBM data on several variants of Stryker vehicles. The exercise drew from a computer database of more than 5 billion records.
“We did a proof of principle at LOGSA using historical maintenance records looking at data to assigned sets of equipment. We matched it to the documentation with technical manuals. We can use AI to see what vehicles look like now and in the future based on miles and sensor reading. AI allows a computer to make those technical connections between engineering specifics and sensor data,” Kuenzli explained.
At the same time, the clear advantages of wireless networking and faster processing speed, somewhat paradoxically, bring the disadvantage of adding security challenges. Real-time wireless transmission is, of course, inherently less secure than merely having sensors gather data to download through wired connections after the passage of time.
“We can’t just go to a Bluetooth or Wi-Fi. We have to make sure the data is secure and protected,” Kuenzli said.
An ability to recognize mission failure more efficiently by leveraging cloud technology and AI to enable faster conditioned-based maintenance, quite understandably, brings significant tactical and financial advantages.
“The return on investment with this will be in the billions and help us prove out that we can reduce the risk of mission failure,” Kuenzli said.
Kris Osborn is a former editor of Defense Systems.