A Suboptimal Fusion Estimation Algorithm Weighted by Matrices Based on LMI and Machine Learning

This article is concerned with suboptimal fusion estimation weighted by matrices for multisensors linear systems with unknown cross-covariance. First, for reasonable correlation coefficients, the simplest constraints are derived by Shure complement theorem. The constraints can ensure the positive definiteness of the extended estimation error covariance matrix and the fusion estimation error covariance matrix weighted by matrices, and the consistency of the proposed fusion estimation. Then, based on the linear matrix inequality (LMI) algorithm combined with AutoML, a suboptimal fusion estimation weighted by matrices is proposed. The proposed suboptimal fusion algorithm provides an effective way for fusion estimation of linear distributed systems with unknown cross-covariance. Simulations analyses verify the effectiveness and correctness of the conclusion.