The Theory of Data Fusion Based on State Optimal Estimation
Wu Yi · Mathematica Applicata · 2007
This paper advances a data fusion method with optimal weighted and analyzes the distribution principle for optimal weighted value;presents an unified linear fusion model for multi-sensor information which can not be restricted by data types and fusion system structure,moreover indicates that it is equivalent between linear minimum square estimation fusion and weighted least squares estimation fusion on premise of positive definite noise covariance matrix;introduces Bayes maximum verified estimation fusion method for data fusion and deduces general expression formula of sensor data fusion using maximum verified method;finally,takes data fusion for two sensors as an example to testify that the fusion state precision obtained by Bayes maximum verified estimation is much higher than that obtained by maximum likelihood estimation under the same condition;but both of them are much better than the local estimation precision for each single senor.