The invention relates to a position coordinate
estimation method based on partial least square regression, and belongs to the technical field of indoor positioning. The method comprises the following steps: S1, constructing initial independent variable and dependent variable matrixes X0 and Y0 by using the position fingerprints, wherein position fingerprints and position coordinates of Nf reference points are known; S2, preprocessing X0 and Y0 to obtain X and Y; S3, constructing an
optimization problem according to the first pair of principal components t1 and u1 of X and Y and the axial vectors w1 and c1, and solving t1 and u1 according to a
covariance maximum principle under certain constraints; S4, performing regression modeling on the principal components t1 and u1; S5, taking residual matrixes E1 and G1 in X and Y as new X and Y, continuing to extract new principal components, and repeating the steps until the number of the principal components reaches an upper limit; S6, based on a PLSR
algorithm, obtaining a regression equation of mapping X to Y; and S7, obtaining position fingerprints of the test points, and then obtaining position coordinates of the test points by using the independent variable
coefficient matrix A in the regression equation. According to the method, the operation complexity is reduced, and meanwhile, the
estimation precision of the position coordinates is ensured.