基于GPU加速弹跳射线法的复杂目标散射特性快速分析

    Fast Analysis of Electromagnetic Scattering from Complex Targets Based on GPU-accelerated SBR Method

    • 摘要: 针对电大尺寸复杂目标的雷达散射特性分析问题,采用GPU加速并结合BVH树结构的弹跳射线法(SBR)计算了复杂目标雷达散射截面的角分布特性。利用C++AMP实现了GPU加速SBR,加速了射线追踪部分,提高了SBR的射线追踪效率。利用BVH树结构加速了射线求交速率,并引入了SAH划分法来划分场景中的面元,避免了包围盒重叠,提升了BVH树构建的准确度和效率。通过与商业电磁软件FEKO中的矩量法(MoM)和射线追踪法(RL-GO)进行对比,说明了文中GPU加速SBR的正确性,且在计算耗时和内存占用方面具有优势。随后分析了载弹和不载弹情形下,复杂飞机目标雷达散射截面的角分布特性,数值结果表明文中算法能够处理电大尺寸复杂目标的雷达散射特性,比传统SBR更具有效性和高效性。

       

      Abstract: Aiming at the analysis of the radar scattering characteristics of electrically large complex targets, the shooting and bouncing rays (SBR) method, which is accelerated by GPU and combined with the BVH tree structure, is used to calculate the angular distribution characteristics of the radar scattering cross-section of a complex target, the angular distribution characteristics of the radar scattering cross section. The GPU-accelerated SBR is implemented using C++AMP, which accelerates the ray tracing part and improves the ray tracing efficiency of the SBR. The ray intersection rate is accelerated using the BVH tree structure, and the SAH division method is introduced to divide the face elements in the scene to avoid the overlapping of the enclosing boxes, which improves the accuracy and efficiency of the BVH tree construction.By comparing with the method of moments (MoM) and ray-tracing method (RL-GO) in the commercial electromagnetic software FEKO, it is demonstrated that the presented GPU-accelerated SBR show good validity and reliability, which also shows advantages in computational time and memory consumption. The angular distribution characteristics of radar scattering cross sections of complex aircraft targets with and without missiles are analyzed and discussed, and the numerical results indicate that the GPU-accelerated SBR in this paper is able to deal with the radar scattering characteristics of electrically large complex targets, and it is more effective and efficient than the traditional ray tracing method.

       

    /

    返回文章
    返回