Specifically, using advanced methods of such as deep learning, NASCTN is investigating detection algorithms for federal incumbent radar signals in the 3550-3650 MHz band. Recent work using spectrograms was reported in a journal paper Deep Learning Classification of 3.5 GHz Band Spectrograms with Applications to Spectrum Sensing and an accompanying technical note Spectrum Occupancy and Ambient Power Distributions for the 3.5 GHz Band Estimated from Observations at Point Loma and Fort Story. The results demonstrated that training deep learning algorithms with real-world measurements can outperform traditional energy detection methods. Moreover, they demonstrated the utility of machine learning for estimating spectrum occupancy and ambient power distributions.
For this and future NASCTN projects involving machine learning, NASCTN has acquired a high-performance workstation (pictured) specifically designed for deep learning applications.