Learning Robust Representations for Automatic Target Recognition
This work addresses the challenge of improving target recognition accuracy in RF systems for defense or surveillance applications, but it appears incremental as it focuses on evaluating existing techniques rather than introducing a new method.
The paper tackled the problem of robust classification for automatic target recognition in radio frequency sensors, particularly against attacks that mask true target characteristics, by evaluating different techniques for building robust classification architectures using learned physical structure in synthetic aperture radar signals of simulated 3D targets, but no concrete results or numbers were provided.
Radio frequency (RF) sensors are used alongside other sensing modalities to provide rich representations of the world. Given the high variability of complex-valued target responses, RF systems are susceptible to attacks masking true target characteristics from accurate identification. In this work, we evaluate different techniques for building robust classification architectures exploiting learned physical structure in received synthetic aperture radar signals of simulated 3D targets.