摘要
针对传统高斯马尔柯夫过程在惯性敏感器随机误差建模过程中存在的由自相关序列计算不准确引起建模不准确的缺点,引入自回归过程建模方法对惯性敏感器的随机误差进行建模。建立了自回归过程模型,分析了三种确定自回归过程模型参数的方法,对三种参数估计方法进行了比较,分析结果表明burg法得到的自回归模型稳定性更好。利用静态数据对随机误差的自回归过程模型进行了测试,并将测试结果与其它随机过程模型得到的结果进行了比较。结果表明采用burg法确定模型参数的自回归过程模型的收敛性好,3阶自回归过程模型就能得到最佳的估计效果。同其它随机过程模型相比,自回归过程模型的收敛速度快,稳定性好。
The conventional Gauss-Markov process has the shortcoming of low precision in modeling the inertial sensors random errors, and this is caused by the inaccurate calculation of autocorrelation sequence. The modeling technique using autoregressive process was introduced to model the inertial sensors stochastic error. Three methods to decide the autoregressive process parameters were analyzed, and the burg method was proved to be the best one by comparing them. The model was tested and compared with other stochastic processes by static data. The result show that the model parameterized with burg method has good convergence, and 3-order model has the optimal estimation effect. AR model has high convergence velocity and stability compared with other stochastic processes.
出处
《中国惯性技术学报》
EI
CSCD
2008年第2期224-227,共4页
Journal of Chinese Inertial Technology
基金
军队重点科研项目基金(KJ06703)
空军工程大学工程学院优秀博士学位论文创新基金资助(BC06003)
关键词
自回归过程
惯性敏感器
随机误差
建模
autoregressive process
inertial sensor
random error
modeling