基于RWTLS的Allan方差算法简化

Allan方差是现在应用最广泛的随机误差辨识方法之一。大量的试验表明Allan方差可有效地分离出导航过程中的多项随机误差，但是Allan方差也有自身的局限性。针对Allan方差在处理大数据量时计算效率低下、辨识度受粗差的影响较大的问题，本文提出了简化Allan方差算法的方案。首先，在确保Allan方差计算准确的前提下，以提高Allan方差计算效率为目的，对Allan方差算法进行简化；然后，利用抗差加权整体最小二乘（RWTLS）模型的迭代算法对简化后的Allan方差辨识结果进行抗差拟合处理；最后，以光纤式惯性测量单元（IMU）为分析对象，设计试验方案对简化后的Allan方差进行验证。

Allan variance simplification algorithm based on RWTLS

Now,Allan variance is the most widely used method for random error identification.Allan variance can effectively separate random errors in the navigation process.But Allan variance also has its own limitations.Allan variance calculation inefficient when dealing with the high-volume.Allan variance is affected by the gross error.This article put forward the solution of simplified Allan variance algorithm.First,a simplified Allan variance algorithm which can both reduce thecalculation burden and keep the accuracy of the results is proposed.Then taking advantage of robust weighted overall least-squares(RWTLS)iterative algorithm of the model of simplified Allan variance identification results are poor resistance fitting processing.Finally with optical fiber type inertial measurement unit(IMU)as the analysis object,experiment scheme to verify this simplified Allan variance is designed.