基于神经网络的希尔伯特变换器设计方法  

Design of Hilbert Convertor Based on Neural Network Technique

作　　者：刘渭清[1] 高晨[1] 余秋菊[1] 张馨予[1] 

机构地区：[1]西安文理学院物理与机械电子工程学院,西安710065

出　　处：《西安文理学院学报（自然科学版）》2013年第4期66-69,共4页Journal of Xi’an University(Natural Science Edition)

摘　　要：提出基于神经最小二乘意义下的数字希尔伯特变换器设计方法.利用Lyapunov能量函数给出希尔伯特变换器幅度误差函数的网络表示形式,当Hopfield网络演变到稳定状态时,即可获得最优的幅度逼近函数,进而得到所需的希尔伯特变换器.Hopfield反馈型神经网络易于采用模拟大规模集成电路实现的特点,保证了设计的实时性.仿真结果表明这种设计是一种高效的、适合硬件实现的数字希尔伯特变换器设计方法.This paper proposes the design of digital Hilbert convertor based on the least squares neural network algorithm. The network representation of amplitude error function of Hilbert Convertor reformulated by the Lyapunov energy function is given. When Hopfield network is stable, the optimal amplitude approximation function is obtained and hence Hilbert convertor. The real time of design is ensured by the convenience of applying simulated large - scale inte- grated circuit of Hopfield feedback neural network. Simulation results indicate that the proposed approach features high effectiveness and is suitable for hardware implementation.

关 键 词：希尔伯特变换器 神经网络 最小二乘法 实时性 幅度响应 

分 类 号：TN713.7[电子电信—电路与系统]