多波束成像声纳阵列稀疏化技术研究

发布时间：2018-07-23 16:49

【摘要】：多波束成像声纳通过在接收端对大规模换能器阵列进行采样,经数字波束形成在一定角度范围内形成均匀密集分布的接收波束,从而实时获取水下目标的二维图像,在海洋资源开发和水下搜救等方面应用广泛。但是大规模换能器阵列带来声纳系统硬件复杂度的提高和成本、体积、功耗的增加。阵列稀疏化技术是解决上述问题的有效途径之一。现有的稀疏阵列设计大部分都是针对单波束情形,约束条件少。而对于成像声纳这类多波束情形下的稀疏阵列设计,尚未得到很好的解决。因此,本文对适用于多波束情形下的阵列稀疏化技术进行研究,具有重要的理论意义和应用价值。主要研究内容与成果如下:(1)研究了基于Farrow结构多波束形成器的阵列稀疏方法,旨在通过得到一组稀疏的加权系数矢量来实现阵元位置的稀疏。采用Farrow结构指向可调波束形成器,结合基于子孔径旋转的多波束形成方法,仅用一组加权系数即可实现90度视野范围内的538个波束形成;构建了基于Farrow结构多波束形成器的阵列稀疏的凸优化模型,仿真和实测数据表明该多波束稀疏方法可以根据设定的门限得到满意解,但是阵列稀疏率不高。(2)研究了基于智能优化算法和凸优化的多波束阵列稀疏方法。将前人提出的基于粒子群算法的单波束阵列稀疏方法拓展到多波束阵列,仿真结果表明该方法寻优能力不佳、计算量过大;针对现有方法的缺陷,提出了一种基于改进的二进制风驱动优化和凸优化的混合算法,同时对阵列的位置和加权矢量进行优化,直到获得满足方向图性能的最小阵元数目的稀疏阵列;仿真和实测数据表明,该算法能够得到满足性能要求的阵列最优权值和阵列布局,无论是寻优性能还是计算效率相比于现有方法都有了较大的提高。(3)研究了基于列空间相关的多波束阵列稀疏方法。考虑到阵元接收信号本身存在很大的相关性,研究了接收阵列导向矩阵的列空间相关模型;引入投影误差,基于每一列在其他列张成的子空间上的投影误差确定阵元的相对冗余,并设计了详细的算法流程;仿真和实测数据表明,基于列空间相关的稀疏方法相比于前两种方法寻优性能有了提升,并且能在同等的主旁瓣性能下达到最优的阵列稀疏率。综上,本文给出的三种稀疏方法各有优缺点。基于Farrow结构多波束形成器的阵列稀疏方法所需加权系数少、计算量小,但是波束主瓣宽度较宽,阵列稀疏率不高;后两种多波束阵列稀疏方法的寻优性能都强于基于Farrow结构多波束形成器的稀疏方法,但是计算时间有所增加。比较而言,基于列空间相关的多波束阵列稀疏方法综合性能最优。
[Abstract]:By sampling a large transducer array at the receiving end, the multi-beam imaging sonar can form a uniform and dense receiving beam through digital beamforming in a certain range of angles, so that the two-dimensional images of underwater targets can be obtained in real time. It is widely used in marine resources development and underwater search and rescue. However, large-scale transducer array brings about the increase of hardware complexity and cost, volume and power consumption of sonar system. Array sparsity is one of the effective ways to solve the above problems. Most of the existing sparse array designs are for single beam cases with few constraints. But the sparse array design of imaging sonar is not well solved. Therefore, it is of great theoretical significance and practical value to study the sparse array technique for multi-beam cases. The main research contents and results are as follows: (1) the array sparsity method based on Farrow structure multi-beam forming is studied in order to obtain a set of sparse weighting coefficient vectors to achieve the sparse location of the array elements. With the Farrow structure pointing adjustable beamformer and the multi-beamforming method based on sub-aperture rotation, 538 beamforming in the range of 90-degree visual field can be realized with only one set of weighting coefficients. A convex optimization model of array sparsity based on Farrow structure is constructed. The simulation and measured data show that the proposed method can obtain satisfactory solutions according to the threshold set. But the sparse rate of the array is not high. (2) A multi-beam array sparse algorithm based on intelligent optimization and convex optimization is studied. The sparse method of single beam array based on particle swarm optimization is extended to multi-beam array. The simulation results show that the algorithm has poor optimization ability and too much computation. A hybrid algorithm based on improved binary wind-driven optimization and convex optimization is proposed. The position and weighted vector of the array are optimized until the sparse array with minimum number of elements satisfying the performance of the pattern is obtained. The simulated and measured data show that the algorithm can obtain the optimal array weight and array layout to meet the performance requirements. Both optimization performance and computational efficiency have been greatly improved compared with the existing methods. (3) Multi-beam array sparse method based on column space correlation is studied. Considering that the received signals of array elements have great correlation, the column space correlation model of the receiving array guidance matrix is studied, and the projection error is introduced. Based on the projection errors of each column in the subspace of other Zhang Cheng columns, the relative redundancy of the array elements is determined, and a detailed algorithm flow is designed. Compared with the previous two methods, the sparse method based on column space can improve the performance and achieve the optimal array sparsity ratio at the same performance of the main sidelobe. In summary, the three sparse methods presented in this paper have their own advantages and disadvantages. The method of array sparsity based on Farrow structure requires less weighting coefficient and less computation, but the width of the main lobe of the beam is wider and the sparse rate of the array is not high. The performance of the latter two multi-beam array sparse methods is better than that of the multi-beam forming method based on Farrow structure, but the computation time is increased. By comparison, the multi-beam array sparse method based on column space correlation is optimal.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TB565.1

【参考文献】