关键词: 满秩分解, 加权最小二乘, 加权观测融合, 集中式融合

Abstract: A weighted measurement fusion estimation algorithm，not constrained by whether each sensor measurement matrix is the same or not or whether measurement noises are correlated or not，is presented by using full-rank decomposition of matrix and weighted least squares theory.It is proved that the estimation result is equivalent to central fusion Kalman estimation result every moment every time，so it has the global optimality，which can obviously reduce computation burden and is convenient for real time use.Two simulation alternatives for a GPS target tracking system verify its functional equivalence，fastness and optimality.

Key words: full-rank decomposition, weighted least squares, weighted measurement fusion, centralized fusion