【摘要】 为进一步提高中低精度器件微机械陀螺的测量精度,采用一种基于动窗平滑的GM(1,1)改进模型对微机械陀螺零输入条件下输出的非平稳随机漂移信号进行建模,去除了多种因素引起的确定性趋势项,并与常用的线性拟合、多项式拟合方法进行了对比研究.对比仿真结果表明,改进的GM(1,1)模型可以有效地去除漂移中的确定性趋势项,较以往效果有一定改善,去除了趋势项的残差序列经验证为平稳、零均值的随机过程,对该序列建立了AR模型,利用AIC信息准则检验了模型适用性,检验结果表明,AR(3)模型为适用模型.对补偿前后的漂移序列用A llan方差方法进行分析,可以有效实现对不同频段内噪声模型的辨识,对比结果表明,补偿后的漂移噪声参数有所减小,噪声信号得到有效抑制.更多还原

【Abstract】 To increase measuring precision of low-mid precision micromechanical gyroscopes,nonstationary random drift signals output under conditions of zero input to the micromechanical gyroscope were modeled using a modified GM(1,1) model.This modeling was based on data smoothing of moving window,eliminating determinate trend terms caused by many factors.The model was compared with common linear fitting and polynomial fitting.Simulation results showed that the modified GM(1,1) model effectively removes drifting determinate trend terms,and the sequence of residual errors that the trend term removed were verified to be stationary and zero-mean.An autoregressive(AR) model for this sequence was established and its applicability was tested with AIC information criterion,showing that the AR(3) model is applicable.Using Allan variance to analyze the drift sequences before and after compensation can effectively recognize noise models at different frequency bands.Contrast results indicated that compensated drift noise parameters were reduced,and noise signals were adequately restrained.更多还原