Booz Allen combines machine learning with standard signal processing techniques to survey the EM spectrum, autonomously identify regions of interest, and characterize signals in real time. Using these techniques, Booz Allen can scale signals analysis expertise against large data volumes and enable opportunistic dynamic spectrum access in 5G radio access networks, which ensures seamless communication and efficient spectrum utilization.

To scale and automate signal discovery in large RF datasets, R.AI.DIO signal processing kit analyzes complex samples from RF receivers, identifies temporal and spectral regions with known and unknown signals, and isolates them for rapid analysis by neural networks. This automation allows analysts to focus on new signals, enhancing their ability to manage and interpret complex RF data.

RF specific emitter identification (SEI), or RF fingerprinting, determines if a message transmission is from a known RF transmitter by comparing the physical characteristics of the received signal with known transmitter characteristics. Unique features may include carrier frequency offset, phase noise, clock jitter, IQ imbalance, and properties of amplifiers, filters, and antennas. SEI has been demonstrated on various signal types, including radar, ADS-B, RFID, Zigbee, Wi-Fi, Bluetooth, and 2G/3G/4G/5G. As an example, Booz Allen leverages R.AI.DIO signal processing kit’s machine learning capabilities into real-time airspace monitoring with SEI features to identify known and unknown aircraft, and aircraft or UAS that may be spoofing their transponder ID. 

R.AI.DIO enhances government organizations’ ability to process growing RF collections efficiently and ensures the speed and accuracy necessary to address modern threats and maintain operational superiority.

Learn how Booz Allen applies AI to signal processing to allow commercial and military users to share the 5G spectrum.