Frequency sharing techniques in a cognitive radar

Abstract

Frequency sharing is the function which allows the Cognitive Radar to perform effective wideband or ultra wideband operations, spanning several frequency channels, by working in parallel with other radar and/or communication systems. The cognitive operation, in presence of concurrent transmitters, is possible by performing two basic processes: modeling of the channel behavior and prediction of the channel occupancy. The model of the electromagnetic environment can be used to predict the future channel occupancy with enough accuracy. This model is based on the observation of the spectrum occupancy during a number of time frames, on the construction of a suitable emulator of the channel behavior and on suitable machine learning of the characteristics of the channel occupancy. This paper describes a complete chain for Frequency Channel Modeling and Prediction, composed of channel observation, channel parameter estimation and channel occupancy prediction, and evaluates the above chain in a typical case study. In order to cope with two main conflicting requirements, namely the large spectrum to be examined and the short time allocated for frequency analysis and prediction, a Compressed Sensing technique is used, in conjunction with Machine Learning for frequency occupancy forecast. We show that, in a typical case study, the use of Machine Learning can ensure a high level of efficiency in presence of a number of concurrent transmitters.