Enhanced Radar Detection and Identification via Compressive Sensing Techniques

Compressive Sensing (CS) is an emerging signal processing theory with a solid mathematical basis, which has rapidly evolved in the recent years becoming a key framework in various areas of applied mathematics, computer science and electrical engineering. It predicts that signals, which allow a sparse representation in some domain, can be recovered from highly incomplete linear measurements (far fewer samples than required by the Shannon-Nyquist sampling theorem) by using efficient techniques. This property has a big potential for practical radar applications as it may accelerate the acquisition speed and reduce operational costs. Moreover, it has been demonstrated that exploiting the inherent sparsity of the radar data can bring further gain in terms of augmented resolution and significant noise and clutter removal. In addition, a dynamic (and eventually knowledge-aided) adaptation of the sampling and processing strategies to the investigated environment results of major interest in the CS application context since adaptive sensing as well as adaptive/learned dictionaries can maximize the acquired amount of relevant information with a reduced number of samples allowing optimum recovery.
The presented work will provide an overview of the research activities of our group and will show some examples of the potential benefits of CS processing techniques in radar target detection and identification problems. These examples cover applications such as beamforming and direction of arrival estimation for arbitrary antenna arrays and MIMO, radar imaging and autofocusing, and novel target classification approaches in diverse scenarios (urban areas, disaster areas, underground, etc.).