论文部分内容阅读
针对运用动态Allan方差提取陀螺随机误差系数时,用截断窗截取原始信号造成方差估计置信度降低的问题,提出运用混合理论方差(Theo H方差)来代替Allan方差对截断窗内的数据进行分析,并提取出随时间变化的陀螺随机误差系数。Theo H方差改善了Allan方差计算时相关时间只能达到信号总时间的二分之一及长相关时间下方差估计置信度降低的问题,其计算的相关时间可以达到数据总时间的四分之三,有效改善了动态算法因数据截取造成误差系数估计置信度下降的缺陷。从对仿真信号和光学陀螺实测数据处理结果上来看,本文方法既能准确地对动态条件下陀螺量测信号的随机误差进行细化辨识,又能大幅提高中、长相关时间下方差估计的置信度。
Aiming at the problem of reducing the confidence of the variance estimation by truncating the original signal with the truncation window when using the dynamic Allan variance to extract the gyroscopic random error coefficients, a hybrid theoretical variance (Theo H variance) instead of the Allan variance is proposed to analyze the data in the truncation window. And extract the gyro random error coefficient changing with time. Theo H variance improves the Allan variance calculation time-related to only one-half of the total signal time and long correlation time lower confidence in the estimation of confidence reduction, the calculation of the relevant time can reach three-fourths of the total data time , Effectively improve the dynamic algorithm due to data interception caused by the error coefficient estimation confidence decline defects. From the processing results of the simulated signals and the measured data of optical gyroscope, the proposed method can not only accurately identify the random errors of the gyro measurement signal under dynamic conditions, but also greatly increase the confidence of the variance estimation under medium-long correlation time degree.