大量的乘法运算和三角函数求值是MUSIC算法实时处理能力不足的主要原因.对MUSIC算法复数乘法进行分解并构造两个实值向量,利用Hermite矩阵的性质去掉冗余乘法运算.在此基础上,提出两种查表法代替三角函数求值:直接查表法,完全避免三角函数求值运算;间接查表法,即对方位角信息和俯仰角信息分别查表,具有较好的信号频率适应性.利用圆阵列结构特征和三角函数对称性减少查表法的数据量,直接查表法仍需较大的存储空间,而间接查表法只需要很小的存储空间.仿真实验结果表明查表法大幅提高了MUSIC算法的计算速率.%The MUSIC algorithm is abundant in the multiplication and the evaluation of trigonometric functions. It is the main reason why it is weak in real-time processing. The complex multiplication of MUSIC algorithm is decomposed and two real vectors are constructed.The redundant multiplication is removed by the properties of Hermite matrix. Based on this,two look-up-table methods are proposed to replace the evaluations of trigonometric functions.One is the direct method, in which all the evaluations of trigonometric functions are avoided.The other is the indirect method, in which the information of azimuth angle and the information of elevation angle are looked up respectively. The latter method has the better adaptability of signal frequency. Using the structural features of circular array and the symmetry properties of trigonometric functions, the data quantity of look-up-table is reduced. The direct method needs much more memory space than the indirect one. Simulation results show that these methods raise the rate of MUSIC algorithm dramatically.
展开▼
