Recent Research on Deep Learning Based Radar Automatic Target Recognition

The focus of this lecture will be recent research and advancement on deep learning (DL) based radar target recognition. Though there is no closed form solution (for target classification) provided by the DL based classification algorithms, there are strategies to achieve a near closed form solution. Analogous to "sufficient statistic and maximum likelihood estimation" our approach here is to define the "sufficient labeled data samples" needed for designing a maximum likelihood classifier. This approach will provide consistent and predictable classification results that we can trust for critical decision making.