Graduating ECE Seniors Michael Zhao and Morriel Kasher have been selected for their Capstone Design project "Radio Frequency Machine Learning (RFML): Streamlining SDR Implementation of Neural Networks Through GNURadio " to compete in the US Airforce Research Laboratory (AFRL) Beyond 5G SDR University Challenge 2021/22, a nation-wide competition that includes both graduate and undergraduate student teams. AFRL also provided funding for the project with high end software defined radio (SDR) equipment that helped significantly in project execution. Michael and Morriel's work advised by ECE Associate Professor Predrag Spasojevic was one of only 11 invitees to the AFRL Beyond Showcase this May 2022.
The RFML project achieves advanced signal processing operations on experimental, real-world radio data in an off-line manner, while post-processing previously collected or simulated data. Application examples include Wideband Anomaly Detection, Radiometric Fingerprinting, and Modulation Recognition, problems of increasing importance due to extreme proliferation of wireless devices. In their work, Michael and Morriel designed and demonstrated a working prototype that can take arbitrary RFML Neural Networks algorithms and signaling formats directly from Software Defined Radios and execute advanced signal processing algorithms in an on-line manner, while the data is received. Their solution is revolutionary in the design and difficulty of achieving and will open many new applications in the field of RFML with very significant implications for current and future wireless systems.
The project was also awarded first place in the ECE Capstone Design competition and third place in the Galbiati Entrepreneurial competition, which recognized those capstone projects with immediate potential for commercialization.
Congratulations to Michael, Morriel and Predrag!