DARPA has tapped BAE Systems to produce a new machine learning-powered tool that allows for the rapid deciphering of unique radio frequency signals, the company said Monday.
The new DARPA deal is worth up to $4.7 million and requires BAE to build a hardware platform that will use the data-driven algorithms it has been developing over the last year for DARPA's radio frequency machine learning systems (RFMLS) program.
"DARPA created the RFMLS program with a focus on the benefits machine learning could bring to signals intelligence," said Josh Niedzwiecki, BAE director of adaptive sensors. "BAE Systems makes a wide variety of SIGINT systems for multiple customers. The CHIMERA system is a reconfigurable front-end that developers can use for prototyping new capabilities that are broadly applicable to that market space."
The new controllable hardware integration for machine learning-enabled real-time adaptivity (CHIMERA) hardware tool is intended to provide a reconfigurable platform that will allow users to decipher radio frequency signals and improve detection of adversaries' jamming and signal disruption attacks.
"CHIMERA brings the flexibility of a software solution to hardware," Dave Logan, BAE Systems' general manager of C4ISR systems, said in a statement. "Machine learning is on the verge of revolutionizing signals intelligence technology, just as it has in other industries."
BAE received a $9.2 million deal from DARPA last November to develop the algorithms that will now be integrated on CHIMERA. The latest contract has covers hardware delivery of the CHIMERA tool, as well as integration and demonstration support. According to Niedzwiecki, it has a three-year performance period ending May 2021.
Nick Zazulia contributed to this article, which was originally published on Defense Daily.