When the B-52 first entered service in the early years of the Cold War, few, if anyone, expected it would go on to serve into the 2020s with talk of additional decades of service. But as programs for replacement aircraft like the B-58 and XB-70 were abandoned, the aircraft found itself outfitted with systems that it’s original engineers, working in the late 1940s, could never have imagined. In another example, Humvees had a planned life of fifteen years but the vehicles are now being modified so that the average fleet age approaches 40 years.
This reality of modern military systems poses a key challenge for Defense Advanced Research Projects Agency (DARPA). Systems are designed for a specific role and then heavily modified in ways the original engineers never predicted and for tasks their systems were never designed for. These improvised adaptations, while necessary, often require skilled technicians to install and their procurements may take years. Moreover, the strain they place on the original system can lead to unexpected system failures and put personnel in danger.
DARPA responded by developing the Learning Introspective Control (LINC) Program. LINC aims to utilize machine learning to enable systems to amend their control laws when faced with unexpected events and to communicate these systems to human and AI operators without negatively impacting continuity if operations or the operators’ confidence.
According to DARPA program manager John-Francis Mergen:
“Today, we start with exquisitely built control systems but then someone needs to add something or make a modification – all of which results in changes to the safe operating limits. […] These changes are done in a way that wasn’t anticipated – or more likely couldn’t have been anticipated – by the original designers. Knowing that military systems will undoubtedly need to be altered, we need greater adaptability. […] When a system ‘wakes up’ in a different space, it needs to be able to realize there are things it can’t do anymore or new things it can, and ‘learn’ how to adapt to its new operating reality. With LINC, we want to provide physical systems with the ability to figure out what is still feasible, alert the operator, and then help them operate in that new space.”
LINC is seeking to achieve its goals in three main research areas.
- “LINC’s first research area will seek to overcome existing limitations in learning models and ML techniques that currently hamper system adaptation.”
- “A second research area will focus on improving how situational awareness and guidance are shared with the operator.”
- “A third research area will focus on testing and evaluating the resulting technologies.”
The agency is hosting a LINC program proposer’s day on 26 August.