Linear minimum variance estimate and optimally weighted LS estimate are often used in many fields such as signal processing,control and communications.

The spatial optimal estimation of groundwater level;

We proposed to integrate the optimal estimation theory with the second-generation wavelet transformation to obtain the wavelet transform coefficients that reflects characteristics of the data.

The optimal estimation of the original measurement information(range,azimuth,elevation,maybe radial velocity)of the.

In order to obtain a wavelet function based on the characteristics of a set of data, an improved algorithm for second generation wavelet transform (SGWT) is proposed by adopting the interpolating subdivision scheme and optimum estimation theorem.

For the difficulty of determining the reliability of each sensor and low precision of observed data,a new fusion method based on supporting matrix and optimal weight is presented.

Through training and simulating,by means of MATLAB and compared with optimal weight filter method,BP neural network provides better accuracy in retraining white noise.

Through training and simulating,by means of MATLAB and compared with optimal weight filter method,BP neural network provides better accuracy in restraining white noise.

This paper brings forward combining these 3 prediction methods after best weighting method,forming a combination prediction method which has the respective advantage of the 3 methods.