Fusion of multi uniform distribution data

Fusion of two and three redundant data with uniform distribution is studied. The minimum square sum of fusion error and the maximum value of fusion error are presented as criteria to evaluate different approaches for the data fusion. A novel fusion method for two data with different accuracy that is called as extended weighted average approach is also presented. The formula to calculate its fusion parameters for two data can be got through theoretic analyze, and the parameters for three data fusion can be got through Monte Carlo calculation. Comparison shows that it is better than three other weighted average approaches include likelihood approach, optimal weighted average and HILARE method. It has the smallest expectation of square sum of fusion error and smallest maximum of the fusion error in all the four approaches. In addition, it is proved that likelihood method and optimal weighted average have the same fusion parameters.