The Improved Diagonal-Matrix-Weight IMM Algorithm Based on LDU Factorization

摘要

The diagonal-matrix-weight IMM (DIMM) algorithm can solve the IMM algorithm confusions of probability density functions (PDFs) and probability masses of stochastic process. However, the DIMM algorithm may generate calculating divergence, caused by the fact that the state error covariance matrix is repeatedly applied to calculate the inverse matrix. A new method. LDU factorization diagonal-matrix-weight IMM (LDU-DIMM) algorithm, is proposed in this paper to solve the factorization of asymmetric matrix error covariance of DIMM algorithm. Experiment results verify the effectiveness of the proposed algorithm.