Cognitive Radar Tracking with Spectrum Sensing and Prediction

This paper considers the problem of target tracking with a cognitive radar that must co-exist in a crowded frequency spectrum with other sources that appear as radio frequency interference (RFI) to the radar. A technique for sensing the spectral usage over time and predicting the spectral occupancy in the next transmission interval is developed based on an alternating renewal process formulation. For each frequency bin in the radar’s available spectrum, the spectral prediction algorithm provides the probability the bin will be occupied and the expected level of RFI in the bin. This information is used by the cognitive radar’s perception-action cycle to choose the next waveform to avoid the RFI and optimize the radar’s detection and tracking performance. A simulation example is provided to demonstrate performance in a randomly time varying RFI environment.