基于神经网络的变换域通信系统接收技术

发布时间：2018-06-01 19:18

本文选题：变换域通信系统 + 基函数　；参考：《华中科技大学学报(自然科学版)》2014年02期

【摘要】：提出了一种不须要估计本地基函数的变换域通信系统接收机接收方法.详细分析了接收机的接收原理,以相关解调信号能量最大化为目标构建了接收机最优化接收模型,通过对模型求解得到了接收数据自协方差矩阵的最大特征值向量,此向量可作为接收端基函数实现最佳接收,从而引入基于Hebbian学习训练的神经网络完成最大特征值向量的自适应估计与数据接收.仿真结果表明:当算法收敛时,作为基函数估计的网络连接强度矢量与发送端基函数基本一致,并且当发送端改变基函数时,算法依然能够实现准确跟踪.系统的接收性能与收发两端基函数一致条件下的接收性能基本一致,当算法迭代次数为300,误码率为1×10-4时,系统的信噪比损失可小于1dB.
[Abstract]:In this paper, a receiver receiving method for a transform domain communication system without estimating the local basis function is proposed. The receiving principle of the receiver is analyzed in detail, and the optimal receiver receiving model is constructed with the aim of maximizing the energy of the correlated demodulation signal. The maximum eigenvalue vector of the self-covariance matrix of the received data is obtained by solving the model. This vector can be used as the basis function of the receiver to realize the optimal reception, and then the neural network based on Hebbian learning training is introduced to realize the adaptive estimation and data reception of the maximum eigenvalue vector. The simulation results show that: when the algorithm converges, the network connection strength vector estimated as the basis function is basically the same as the transmission terminal basis function, and when the transmitter changes the basis function, the algorithm can still achieve accurate tracking. The receiver performance of the system is basically the same as that under the condition of consistent transceiver and transceiver basis functions. When the iterative number of the algorithm is 300 and the bit error rate is 1 脳 10 ~ (-4), the SNR loss of the system can be less than 1 dB.
【作者单位】：空军工程大学信息与导航学院;中国人民解放94900部队;
【基金】：国家自然科学基金资助项目(60972042,61271250) 空军工程大学信息与导航学院研究生论文创新基金项目(2011004)
【分类号】：TN92;TP183

【相似文献】