Abstract:

To overcome the lower data fusion precision, which is caused by the restriction of noise measurement to dissimilarsensor, a novel data fusion method based on weighted least square is proposed. The multisensor measurement model is analyzed. Then the weighted fusion formula is derived for parameter fusion estimation from multiple dissimilarsensors, which is based on the principle of least square; and the scaling coefficients of weighted fusion of various sensors and estimated variances of erroneous vectors are obtained. Coparing with the error of singlesensor and mean estimation, the measurement accuracy is higher for dissimilarsensor data fusion based on weighted least square. Numerical case result is in consistent with theory deduction that further validates the proposed methodology.

Key words: information processing theology, dissimilarsensor, data fusion, weighted least square