Improving compressive sensing results in radar using multiple reconstructions

Compressive sensing based reconstruction introduces noise which is dependent on a number of factors, in particular the choice of representation basis. In this paper we show how multiple reconstructions using different bases can be used to more accurately retrieve target information in a radar signal. We focus on signal averaging as a technique for achieving these improvements, and discuss the effectiveness of this strategy as well as a few potential problems and limitations inherent in such a strategy. We also provide a basic example of a way of improving this averaging technique, and provide a template for further development and case-by-case fine tuning.