摘要
提出了一种新的基于高阶累积量和神经网络的盲波束形成器,并对其进行了建模与仿真研究。该波束形成器应用高阶累积量对期望信号的方向向量进行估计,采用线性规划神经网络实现盲波束形成,不但减少了对阵列流形的依赖,具有较好的容差性,而且能有效避免矩阵求逆运算,减小了运算量,易于用硬件实时实现。仿真实验验证了该波束形成器模型的有效性和正确性。
A new blind beamformer based on higher-order cumulant and neural network is presented in this paper. Its model and simulation are studied further. In this beamformer, the higher-order cumulant is used to estimate the steering vector of the desired signal and the linear programming neural network is employed to carry out the blind optimizing beamforming. The algorithm is not very dependent on the manifold of the array and adapts to the error easily. It is able to avoid computing the inverse of a matrix so as to cut down on computing amount and be easily implemented by hardware. Simulation proves efficiency and correctness of this beamformer model.
出处
《系统仿真学报》
CAS
CSCD
2002年第8期983-986,共4页
Journal of System Simulation
基金
国防科技重点实验室基金项目(编号:2000JS23.2.1)
船舶国防科技预研基金项目(编号:2000J42.2.8)。
