Multi-dimensional-parameter estimation method of optimal fusion of multiple types of heterogeneous models

Abstract

The invention discloses a multi-dimensional-parameter estimation method of optimal fusion of multiple types of heterogeneous models. The method includes: acquiring M types of heterogeneous observationdata, and determining multi-dimensional parameters; determining model relationships between the M types of heterogeneous observation data and multi-dimensional parameters to obtain the M types of heterogeneous observation models; and constructing a fusion optimization function, and solving the fusion optimization function to calculate optimal weighting values of the M types of heterogeneous observation models and estimation values of to-be-estimated multi-dimensional parameters. In carrying out fusion processing of the multi-dimensional heterogeneous unequal-precision data through calculationof estimation deviations and an estimation mean-square-error matrix of the to-be-estimated multi-dimensional parameters in the method for nonlinear models, the optimal fusion weight values are no longer determined solely by precision of the measurement data, but are related to model structures at the same time, a calculation method of the optimal fusion weight values in actual data fusion processing is given, at the same time, a corresponding multi-dimensional parameter estimation algorithm is established, and an effect that the optimal fusion weight values can be further precisely calculatedand obtained in data fusion processing is realized.

Description

technical field [0001] The invention relates to the technical field of data fusion, in particular to a multi-dimensional parameter estimation method for optimal fusion of multi-class heterogeneous models. Background technique [0002] In measurement data fusion processing, the most typical is the fusion of multi-source heterogeneous and unequal precision data. After expressing the observed data as a parameter model, the measurement data fusion problem can be transformed into the parameter estimation problem of the regression model. The fusion processing of measurement data is beneficial to improve the accuracy of parameter estimation. Among them, different types of data or heterogeneous data means that the functional relationship of the observed data with respect to the parameters to be estimated is different, so that the derivatives of each order are also different. If the fusion process involves heterogeneous and unequal precision observation data, the different weightin...

AI Technical Summary

Problems solved by technology

[0005] However, in the current fusion process of multi-type heterogeneous observation data, it is often assumed that all observation data are of equal precision, that is, the random errors of observation data are independent and identically distributed, so weighting is not considered, or the Gauss-Markov theorem of linear models is used directly Conclusion Weighting multi-category heterogeneous measurement data reduces the accuracy of data fusion

Images