|
|||||
|
|
| 基于神经网络的无源多传感器属性数据关联 | |
| 引用本文： | 徐敬,王秀坤,胡家升. 基于神经网络的无源多传感器属性数据关联[J]. 系统仿真学报, 2003, 15(1): 127-128,131 |
| 作者姓名： | 徐敬 王秀坤 胡家升 |
| 作者单位： | 1. 大连理工大学,大连,116024;海军大连舰艇学院,大连,116018 2. 大连理工大学,大连,116024 |
| 摘 要： | 采用引入动量项、自适应调整步长，Levenberg-Marquardt优化方法对基本的BP神经网络进行改进，以提高学习速度，改进的BP神经网络学习算法用于对无源多传感器获得的雷达辐射源参数进行属性数据关联，能够自适应地调整阈值，即根据训练数据调整关联的门限值，与确定门限的属性关联算法相比，有着很高的关联正确率。 |
| 关 键 词： | 神经网络 无源多传感器 属性 数据关联 数据处理 阈值 BP算法 学习算法 |
| 文章编号： | 1004-731X(2003)01-0127-02 |
Multiple Passive Sensors Feature Data Association Based on Neural Networks |
|
| XU Jing,,WANG Xiu-kun,HU Jia-sheng. Multiple Passive Sensors Feature Data Association Based on Neural Networks[J]. Journal of System Simulation, 2003, 15(1): 127-128,131 | |
| Authors: | XU Jing WANG Xiu-kun HU Jia-sheng |
| Affiliation: | XU Jing1,2,WANG Xiu-kun1,HU Jia-sheng1 |
| Abstract: | An introduced momentum item, adaptive step adjust and Levenberg-Marquardt optimal methods are used to improve the basic BP neural networks, and training speed is highly developed as well. The improved BP neural networks learning algorithm is presented to associate the feature data of radar emitters received by multiple passive sensors. It can adapively adjust the threshold, i.e. adjust the feature associate threshold according to training data. Compared with fixed threshold feature association algorithm, it shows a high association rate in feature association. |
| Keywords: | feature data association BP neural networks improved BP learning algorithm threshold |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |