Air Force Wants 'Data Science Ecosystems' to Shorten OODA Loops
The AF Research Lab is looking for ways to deal with the rising ocean of surveillance video and other sensor data.
The Air Force’s military superiority is centered around the service and individual airmen’s ability to ingest new information and act quickly and decisively. But the unending waves of data being collected through modern methods have made that harder, and top Air Force scientists are looking for new ways to streamline the process.
In the 1950s, Air Force Col. John Boyd [no relation to this article’s author] developed a method for systematizing the decision-making process for combat pilots: the OODA Loop, which stands for observe, orient, decide and act. According to Boyd—and decades of Air Force and military strategy since—the combatant with the shortest OODA Loop will have the advantage in any fight.
As the Air Force continues to rapidly expand the amount of data it collects, the service is focused on ingesting and processing that information in a timely fashion. To that end, officials at the Air Force Research Lab Information Directorate are looking to create an Air Operations Center “Data Scientist’s Ecosystem” that can take in unstructured data and clean it up for use by Air Force data scientists and automated tools.
“The Air Force is interested in novel methods to develop a work bench, work flow, or ecosystem that enables rapid integration of new data for use in operations centers and brings it to bear on mission directives to improve the efficiency and accuracy of operations,” according to a request for information posted Monday to beta.SAM.gov.
“Often the processing necessary to make data available in operations centers takes so long that the operational relevance of the data expires,” defying the basic tenets of the OODA Loop, the RFI states. “Expanding our ability to rapidly process and leverage new data while it’s still relevant is required in order to provide our commanders with decision-advantage.”
The document notes the current structure for ingesting and analyzing data puts the burden on the operator—someone who should be focused on moving through the cycle toward action and, often, has little to no data science background. The plan to remedy this situation is two-fold: embed data scientists in operations centers and develop tools and workflows to enable a data-centric decision-making ecosystem.
“This ecosystem would be used by future data scientists as the USAF transitions to Joint All-Domain Command and Control,” the RFI states, “and would provide commanders decision-advantage over our adversaries by making data and analytics accessible while the data is still operationally relevant.”
The RFI asks for feedback on “technical concepts, approaches, ideas, algorithms, research, prototypes and merits of the ideas of work” for six focus areas:
- Problems in the OODA Loop cycle that data science could help solve.
- Analytic use cases that would demonstrate the value of data science in the Air Operations Center.
- Techniques that help eliminate manual data wrangling.
- Promoting and casting mission relevant data and intelligence to machine readable formats.
- Fusion, integration and alignment of disparate data sources,
- Mapping data to the intended use through analytics and visualizations.
The RFI includes additional details on each of the six priority areas.
Contracting officials added they are looking for technologies and approaches that can be integrated into the existing Air Operations Center architecture and have no plans to upgrade that infrastructure as part of this effort.
The Air Force is also using the RFI to gather information on the best acquisition strategy, including possibly establishing a long-term broad agency announcement, or BAA, that can be used as a base for future solicitations.
Responses are due by October 26.