A Comparison of cognitive approaches for clutter-distribution identification in nonstationary environments

Most existing radar algorithms are developed under the assumption that the environment (clutter) is stationary. However, in practice, the statistical characteristics of the clutter can vary enormously in space, time, or both, depending on the radar-operational scenarios. If unaccounted for, these nonstationary variabilities may drastically hinder the radar performance. Therefore, to overcome such shortcomings, the cognitive radar framework is being developed to dynamically detect changes in the clutter characteristics, and to adapt to these changes by identifying the new clutter distribution. In this work, we present a sparse recovery based clutter identification technique, and compare its performance with the Ozturk algorithm based clutter identification method. The sparse recovery based technique uses kernel density estimation method to create the dictionary, and applies the batch orthogonal matching pursuit algorithm to identify the clutter distribution. With numerical examples we demonstrate that, in comparison to the Ozturk algorithm based method, the sparse recovery based technique provides (i) improved accuracy in identifying clutter distributions that have different parameters, but are from the same family; and (ii) robustness in terms of measurements used for dictionary generation and test distribution identification.