Waveform Adaptation for Target Classification Using HRRP in a Cognitive Framework

Waveform adaptation is a key-feature for modern radar systems and essential for cognitive radar. In this article, we present a concept for the enhancement of the classification performance by using optimized transmit waveforms and a Gaussian template matching high range resolution profile classifier. A straight forward approach is presented, aiming to improve specific parts of the confusion matrix, which will be exploited within a cognitive framework. The optimization includes different types of uncertainties and is designed during a training process to be accessed by a library. Taking different uncertainties into account, the calculation of the expected performance, the optimization, the range side lobe constraint and the time-domain realisation is explained. A nonlinear frequency modulation waveform is used since it provides a compression gain with range resolution and a constant envelope. Based on an electromagnetic simulation, the concept is validated for different ground targets and aspect angle uncertainties. The adaptation is compared to a commonly used linear frequency modulation. The results of the mean performance improvement reached an enhancement between
$8.8 \,\%$
and
$20.9 \,\%$
.