Design of Vision Transformer for Range-Doppler Map-Based Radar Target Classification

Range-Doppler map is the principal information that a radar system can acquire from an antenna that is illuminating the target. Radar system can extract the relative range, relative velocity, and signal strength of the target from a range-Doppler map. The distribution of the signal strength varies according to the target shape. Using these characteristics, the radar system can distinguish between the target types. This paper proposes the design of vision transformer-based deep learning network for target classification from a range-Doppler map.