论文部分内容阅读
为解决无线分集相干光接收机的自适应盲检测问题,提出了一种新的离散时间连续状态的网络输出反馈偏置型的复Hopfield神经网络用以解决多值QAM信号的盲检测问题。反馈电压偏置的引入即不脱离传统Hopfield模型,又能有效满足多值信号检测时所需的搜索空间变大的特殊要求。全文完成多值信号盲检测的优化问题构造和能量函数的映射,给出能量函数的证明、分析和它的约束条件,给出适用该问题的激活函数的基本特征,正确盲检测信号的权矩阵的配置方法。最后,通过详细的仿真结果展示和与其他算法性能对比进一步验证算法的有效性和优越性并指出算法所存在的问题和下一步的研究方向。
In order to solve the adaptive blind detection problem of wireless diversity coherent optical receiver, a new Hopfield neural network with output feedback biased discrete-time continuous state is proposed to solve the blind detection problem of multi-valued QAM signals. The introduction of the feedback voltage offset can not meet the traditional Hopfield model effectively and meet the special requirements of larger search space for multi-valued signal detection. In this paper, the construction of the optimization problem and the mapping of the energy function for the blind detection of multi-valued signals are completed. The proof and analysis of the energy function and its constraints are given. The basic characteristics of the activation function for the problem, the weight matrix of the blind detection signal The configuration method. Finally, the effectiveness and superiority of the proposed algorithm are further demonstrated through detailed simulation results and comparison with other algorithms. The problems existing in the algorithm and the research direction are also pointed out.