Chapter 6: Sparsity-based radar technique

The output of a radar front-end is typically a vast data stream which contains only a few parameters of interest, e.g., for a directional antenna array the specific azimuth angles of arrival for signal sources. If the observed scene can be considered to be sparsely populated compressive sensing or sparse recovering techniques allow to reduce the required number of samples and/or sensors without degrading the performance of the system. This leads to a dramatically cut-down of the acquired data-stream which is accompanied by a considerable cost reduction of the overall system. In many radar applications the observation scene can be considered to contain only a limited number of objects, which makes the sparse reconstruction techniques very attractive to implement them in these systems. The main feature of this new technique is the potential to estimate accurately target states from a few elementary measurements selected from a fixed Nyquist-Shannon sampled environment. Due to this fact the vast data stream can be reduced dramatically. This paper provides the interested reader with a basic understanding of the power of sparse reconstruction techniques and its application to radar or sonar systems. A few properties of sparse reconstruction algorithms in the area of radar and sensor networks will be discussed throughout this article and several examples are presented to prove the concept.