高级检索

二维Root-MUSIC算法的快速实现方法

Research on fast-realization of 2D Root-MUSIC algorithm

  • 摘要: 目前实际测向系统中所需DSP数目较多,系统设计较复杂,测向时效性不足。针对于此,基于一种改进的二维DOA估计算法—2D FD-Root-MUSIC算法,研究了在FPGA+DSP平台上快速实现DOA估计。首先介绍了数据的量化误差及分析,给出了算法的定点编程实现流程,然后对浮点数据进行量化仿真,确定算法在硬件中的定点位数,最后通过对浮点数据进行量化仿真的结果,设计了协方差矩阵、噪声子空间、谱值及FFT的定点计算结构,并分析了各单元运算速度及资源量。仿真结果表明,改进算法经过并行化处理后,实现速度可有效加快。

     

    Abstract: In modern DOA system,a large number of DSPs are needed and the system design is very complexity and its real-time performance is not very good.Because of this,an advanced algorithm for realizing 2-D DOA(2D FD-Root-MUSIC) based on FPGA and DSP is studied in this paper.Firstly,the problem of quantification errors and cal-culation errors is anayzed before the algorithm is realized on hardware platform,and the flow of fixed-point simulation is introduced.Then,the number of the fixed-point data is fixed on quantizing float data simulation.Finally,according to the result of quantized float data simulation,the fixed-point structure of covariance matrix,noise subspace,spatial spectrum value and FFT are designed.The speeds of calculating and resources needed are also analyzed.Simulation re-sults show that the advanced algorithm can be realized more quickly and efficiently after parallel processing.

     

/

返回文章
返回