Regression imputation under incomplete auxiliary information and its variance estimate;

In this peper, it is shown that combined uncertainties(or variance estimates) in indirect measurements have different number of degree of freedom to independent and dependent direct measured quantities.

For missing data,based on the unbiased estimation of the population quantity of the non-responser of the interest variables and the auxiliary variables,regression imputation using response probability and its variance estimate are formulated by resampling technologies such as Jackknife and Bootstrap.

A study of variance estimation of kalman filtering method;

A new method of variance estimation based on wavelet packet de-noising;

Dynamic compensation based on noise variance estimation and model reference for sensors

In order to solve the disadvantage that the variance-estimation algorithm based on the constant window length can not follow noise changes accurately in multi-sensor data fusion,a new variance-estimation algorithm is proposed,based on alternating the window length according to noise changes.

The validity,accuracy and actual time of the algorithm for batched-estimation,self-adaptive weighting and variance-estimation are studied in multi-sensors data fusion.

Subsequently the estimation consistency is proved,and the estimation variance is derived theoretically.