U.S. Looks to Smart Radio to Secure Data Networks
U.S. military and intelligence agencies are looking for ways to boost data security when communicating “in the wild.”
Among the proposed approaches are emerging smart radio platforms that can secure classified information. Heading one effort is the Intelligence Advanced Research Projects Activity (IARPA), which is soliciting industry proposals under a program dubbed Securing Compartmented Information with Smart Radio Systems, or SCISRS.
The program seeks to plug security gaps between sensitive government facilities where data is transmitted, received and stored. For example, transmissions within a sprawling U.S. Navy base could be breached via unexpected RF transmissions.
“The goal of the SCISRS program is to develop smart radio techniques to automatically detect and characterize these suspicious signals and other RF anomalies in complex RF environments,” explained Paul Kolb, IARPA’s program manager. Anomalous signals are often used to mask exploits intended to breach RF communications links.
Examples of exploits range from unexpected transmissions at a particular location “or signals that don’t otherwise belong,” said Kolb. “A signal mimicking a cell tower, for instance, that’s not a good sign. It means something might be trying to communicate with your device.”
Among the possible security solutions is using “advanced smart radio techniques to automatically detect and characterize RF anomalies,” he added.
Among the potential data security solutions are radio frequency machine learning, software-defined radios, security platforms that combine GPUs and CPUs and “classical” signals analysis, Kolb told a recent “Proposers’ Day” gathering held by IARPA.
The agency said emerging software-defined radios and published data sets have advanced that state of the art for smart radio technology. “What once took expensive, specialized equipment can now be done with an SDR and a laptop,” Kolb noted. Add to that, large data sets “have opened [RF machine learning] research to researchers without an RF lab.”
The smart radio effort also builds upon earlier RF machine learning work at the Defense Advanced Research Projects Agency, Kolb added.
In selecting smart radio suppliers, the agency said it looking at a combination of machine learning and signal processing techniques. Those range from unsupervised and supervised learning and the combination of neural networks and deep learning to signal recovery and filtering techniques.
The deadline for industry proposals was Nov. 13. The agency expects to award contracts for smart radio development early next year, with the initial design reviews 18 months hence. Program officials said IARPA is seeking “complementary capabilities” as it looks for industry teams to develop a smart radio infrastructure.
