论文部分内容阅读
为解决无线分集相干光接收机的电域自适应盲均衡的特定问题,根据动力学驱动反馈神经网络(DDRNN)的特点,提出一种基于同相部分与正交部分振幅激励的DDRNN盲均衡方法。设计了一种结构简单、灵活性强的连续多阈值的正弦函数型激励函数形式,详细分析了该激励函数的参数选择方法并提出了该问题接近点和远离点的新概念;进而从激励函数角度分析了DDRNN电域自适应盲均衡问题的放大因子取值。论述了高阶正交振幅调制(QAM)的激励函数用于低阶QAM系统盲均衡时的适用性。设计了多阈值激励函数和DDRNN新的能量函数并给出相关证明,该激励函数和能量函数的设计和分析不仅可适用于无线相干光接收机的自适应盲均衡,也可推广至其他相关领域。
In order to solve the problem of domain-adaptive blind equalization in wireless diversity coherent optical receivers, a DDRNN blind equalization method based on amplitude-driven in-phase and quadrature parts is proposed according to the characteristics of the dynamics-driven feedback neural network (DDRNN). A simple and flexible multi-threshold sine function excitation function is designed. The parameter selection method of the excitation function is analyzed in detail and a new concept of approach point and farther away from the point is proposed. From the excitation function From the point of view, the value of the amplification factor of adaptive blind equalization problem in DDRNN domain is analyzed. The applicability of high order Quadrature Amplitude Modulation (QAM) excitation function for blind equalization of low order QAM systems is discussed. The multi-threshold excitation function and DDRNN new energy function are designed and the related proofs are given. The design and analysis of the excitation function and energy function are not only applicable to adaptive blind equalization of wireless coherent optical receivers, but also to other related fields .