RARSP: Rapid and Accurate Radar Signature Prediction

Modeling of radar signature of sea targets in dynamic sea states is a critically important problem in developing methods of detection and identification of potentially threatening ships. As most maritime radars operate at X-band, this EM problem has an extremely large electric-size and it is further complicated by the sea wave phenomena. Simulation tools exist for high-frequency electromagnetic (EM) simulation. However, existing tools are insufficient in following three aspects: incapable of modeling the fine features on the topside of ships which often have significant scattering contributions due to their comparable size to X-band wavelength; incapable of capturing the interaction between ships and complex sea states; not suitable for state-of-the-art Graphic Processing Unit (GPU) or GPU-cluster acceleration. We propose to develop a hybrid method based on the novel Bidirectional Analytic Ray Tracing (BART) algorithm and the 3D fast Method of Moments (MoM) algorithm. Besides the fine features of ships, the proposed tool can also take account of scattering of rough sea surfaces. Both BART and MoM can be accelerated by inexpensive GPUs.