US Patent 12028095 Machine-learning based analysis and response to electromagnetic waveforms

A method includes determining, based at least in part on parameters of a software-defined radio (SDR), waveform data descriptive of an electromagnetic waveform. The method also includes generating feature data based on the waveform data and based on one or more symbols decoded from the electromagnetic waveform. The method further includes providing the feature data as input to a first machine-learning model and initiating a response action based on an output of the first machine-learning model.