【摘要】 实时识别陀螺漂移,并且预测漂移性能的变化趋势,对提高整个导航系统的精度有着十分重要的意义。本文结合小波分析理论和小波神经网络的非线性预测算法,对陀螺信号进行趋势提取与漂移预测。采用小波神经网络避免了其它神经网络存在的局部最小化的缺陷,小波分析的引入可以有效提取出原信号的趋势,大大降低了环境因素的影响。将小波神经网络非线性预测算法与小波趋势提取算法结合,建立陀螺仪漂移趋势的预测模型,对某光纤陀螺实测信号仿真预测其漂移趋势,仿真结果证实该预测模型与实际情况相符,具有较好的预测精度,为预测一般陀螺的随机漂移提供了一种新的有效途径,同时还对一般陀螺仪表的漂移模型建立方法提供借鉴。更多还原

【Abstract】 Discerning gyro’s drift in real time,and predicting the variation tendency of gyro’ s drifts are very important to improve the accuracy of the whole navigation system.In this paper,wavelet analysis theory and non-linear wavelet neural network prediction algorithm are adopted for the gyro’s drift extraction and prediction.The wavelet neural network is adopted to prevent the local minimize defects that exits in other neural networks,and the wavelet analysis is introduced to extract the trend of signal effectively which could greatly reduce the impaction of environmental factors.Wavelet neural network nonlinear prediction algorithm combined with wavelet extraction algorithms which can set up the prediction model and predict the trend of drifts for measured gyro signal.The simulation results showed that the prediction model in line with the actual situation.This method provides a new ways for forecasting the gyro’s drift signal which include yawp signal which can improve the precision obviously;at the same time,the method could provide the reference for modeling the general gyro’s drifts.更多还原