基于TH神经网络的自适应滤波在船舶姿态测量信号处理中的应用研究

自适应滤波是提高滤波性能的主要方法之一。自适应滤波要求实时跟踪输入信号的变化，实时地计算滤波器的权系数。这就大大增加了运算量，因此很难用基于数字计算的方法给出所希望的实时解。神经网络所具有的高度运算能力为解决这一问题展示了光明前景。中利用根据线性规划神经网络导出的TH网络，就船舶姿态测量信号进行滤波计算，仿真结果表明基于TH神经网络的自适应滤波是有效的。

Application of Adaptive Filter of TH ANN in Ship's Attitude Signal Processing

The adaptive filter is one of the methods to improve filtering performance.Because it is necessary for an adaptive filter to trace the change of an input signal and to calculate its realtime weighting coefficients in real time,the calculating work is great.Therefore it is difficult to get a deserable realtime solution by the numerical method.The great calculating capability of artificial neural network ( ANN) cans olve this problem.This paper uses a TH neural network derived by a linear program neural network to filter ship attitude signals.The results of simulation show that the adaptive filter based on TH ANN is efficient.