IS4S & Virginia Tech Co-Host AOC Webinar on Modular Radio Frequency Machine Learning
Chris Bieber, IS4S vice president and site lead at our Aberdeen location, co-presented an Association of Old Crows webinar about modular radio frequency machine learning (RFML) with Dr. William Headley from the Virgina Tech National Security Insititute (VTNSI).
Chris, who established the IS4S Aberdeen site after serving over 20 years in the U.S. Army, and his team of engineers have recently pioneered a modular approach to cognitive electromagnetic warfare (CogEW) along with several partners, including Dr. Headley at VTNSI. Their webinar, titled “Modular RFML: Anomaly Detection and Uncertainty Quantification in Electromagnetic Warfare,” covered what modular RFML is, an overview of the IS4S team’s groundbreaking research, and how it can be used to solve some of the issues facing our military today.
Chris set the stage by describing how congested the RF spectrum has become and how challenging it is to correctly identify what is happening in any given environment. This is a problem not only for the warfighter, but for the intelligence community (the experts that make the official determination of a signal). He shared how his background as an infantryman provides a rare operational view for an engineer and impacts his approach to problems. Thinking like an infantryman, Chris recommended starting from the operational perspective to design systems and then iteratively developing a technical approach to solutioning based off the operational need first.
As the real-world threat environment evolves, one of the major obstacles is the rarity of electromagnetic warfare (EW) expertise across Army formations. This creates a major talent gap for human experts, and an even greater challenge in applying their expertise and knowledge at scale. Chris frames the challenge as how do we take the knowledge of global experts and put access to this expertise in the hands of our soldiers, sailors, airmen, and marines when they need it most? He outlined how one such solution is radiofrequency machine learning. RFML can play a key component to proliferation and automation of these expert skills to help meet the immediate needs of our Nation’s finest.
Dr. Headley defined RFML as “the use of state-of-the-art AIML for radio frequency wireless communication data” and highlighted how the majority of AIML research up to this point has focused on modalities outside of RF, and that now researchers are working to catch-up in the RF space. Together, Chris and Dr. Headley reviewed several topics: applications of RFML, approaches to spectrum sensing, approaches to anomaly detection, cognitive modularity components, hardware form factors, and more.
Although limited in the details they could share in a public forum, Chris described how they’ve applied the principals of a modular RFML framework into actual interfaces and the benefits this approach offers. IS4S is developing a modular framework to address challenging EW problems in a manner that enables scalability and freedom to integrate various algorithms and datasets based on mission need.
In summarizing their implementations of RFML, Chris explained how his team is not developing “a solution looking for a problem.” They are focused exclusively on solving real operational problems and getting them to warfighters when they need them, no matter the system constraint or specific EW domain area. The warfighter is the focus, and they are always looking for (and anticipating) their problems that need to be solved.
Both Chris and Dr. Headley encouraged attendees with additional questions or collaborative opportunities to start a conversation.
