Hopfield优化网络用于DOA估计存在的问题
SIMPLE ANALYSIS OF THE NEURAL OPTIMIZATION METHOD FOR DOA ESTIMATION
-
摘要: 本文讨论了用Hopfield优化网络解决DOA估计问题方法的有效性和正确性。虽然这种方法可以避免特征分解和谱峰搜索运算,但在不限制网络中输出为1的神经元个数的条件下,方法中代价函数的构造是不正确的。理论分析和仿真实验都证实了上述结论。
-
关键词:
- 神经网络; 波达方向估计; 谱估计
Abstract: This paper gives a simple analysis of the method of using the Hop-field s optimization neural network to solve the DOA estimation problem. Although the method can avoid the eigendecomposition of data autocorrelation matrix and the orthogonality search of parameter space, theoretical analysis and simulation results show that the construction of the DOA cost function is incorrect on the condition that there is no constraint on the number of outputs of the network. -
Schmidt R O. IEEE Trans. on AP, 1986, Ap-34(3): 276-280.[2]Stoica P, Nehorai A. IEEE Trans. on ASSP, 1989, ASSP-37(5): 720-741.[3]Stoica P, Nehorai A. MUSIC[4]Maximum Likelihood, and Cramer-Rao Bound: Further Results[5]and Comparisons, IEEE Proc. ASSP, 1989, 2605-2608.[6]Kung S Y, Diamantaras K I. A Neural Network Learning Algorithm for Adaptive Principal Component Extraction (APEX), Proc. ICASSP, 1990, Vol. 2, 861-861.[7]Lin Sheng, Yin Qin-Ye. Adaptive Principal Component Analysis Via Neural Networks. IEEE[8]International Conference on Signal Processing, China: 1993, 1224-1227.[9]Jha S, Durrani T. IEEE Trans. on SMC, 1991, SMC-2I(5): 1192-1201.[10]Sharman K C. Maximum Likelihood Parameter Estimation by Simulated Annealing. Proc. ICASSP. 1988, 2741-2744.[11]林盛,殷勤业.信号处理学报,1994,10(2): 66-74.[12]林盛,殷勤业.电子科学学刊,1994,17(1): 1-6. -
计量
- 文章访问数: 1956
- HTML全文浏览量: 88
- PDF下载量: 515
- 被引次数: 0