Accurate and Efficient Evaluation of Spatial Electromagnetic Responses of Large Scale Targets

Abstract

An error controllable algorithm is proposed to efficiently evaluate two-dimensional (2-D) monostatic radar cross section (RCS) from electrically large and complex targets. The algorithm employs interpolative decomposition (ID) to conduct lowrank decomposition on the excitation matrix consisting of multiple right-hand-sides (RHS's) to figure out the so-called skeleton incidents. After the solutions corresponding to skeletons are obtained, the complete angular responses can be recovered efficiently. The proposed algorithm is error controllable because the accuracy/resolution of the excitation matrix and ID can both be manipulated. The algorithm is efficient in terms of CPU time due to the high efficiency of ID. For large scale problems, the number of unknowns and that of incidents become large, which, in turn, would lead to a huge excitation matrix. A multiinterval variation is proposed to overcome the possible bottle-neck on memory usage. The proposed method is user-friendly because all the tunable parameters are problem independent. Numerical experiments on electrically large and complex targets have been performed to show the performance of the proposed algorithm.

Efficient and Accurate Electromagnetic Angular Sweeping of Rough Surfaces by MPI Parallel Randomized Low-Rank Decomposition

“…To accelerate the iterative solution, an algebraic preconditioner [43], [44] can be employed. It has been shown in [25] that the abovementioned skeletonization scheme was strictly controlled by a threshold during the factorization when applying to the plane wave sweeping of deterministic target. As it is known, due to the employing of the tapered wave, the infinite rough surface is truncated into a finite one.…”

“…With the estimated l, the low-rank decomposition can be then carried out on a matrix of size l × n by the standard algorithm, i.e., pivoted QR in this work. From previous studies [25], [39], it is clear that the skeletonization is very efficient in terms of CPU time but requires about O(mn) memory space. This highlights the need for developing parallel randomized matrix decomposition on distributed memory platforms that can meet the memory requirement.…”

“…Many efforts have been devoted to reduce C d [22]- [24] as well as N mc . However, to the best of the authors' knowledge, seldom can be found in terms of n. Encouraged by the angular sweeping algorithm first proposed in [25], this work presents our attempt to improve the efficiency of the MC simulation on large-scale rough surfaces where solve each single deterministic computation. Although the skeletonization has been proved efficient in angular sweeping in terms of plane waves and Gaussian beams, its capability in accelerating the simulation in terms of tapered waves is not well documented.…”

“…Essentially, the skeletonization approach proposed in [25] is based on randomized low-rank matrix decomposition [26]- [28]. Suppose N unk is the number of unknowns for the deterministic computations, V is a right-hand-side (RHS) matrix of size N unk × n, consisting of n RHS vectors.…”

“…The peak memory requirement of the randomized decomposition can be as high as O(N unk • n). Such a memory requirement can hardly be satisfied on a single computer/server if N unk • n is large [25], [29], [30]. Compared with shared memory computers [31], distributed memory computing platforms are more This work is licensed under a Creative Commons Attribution 4.0 License.…”