Multi-sensor data fusion with correlated measurement noises

By using the Cholesky factorization and inverse method for unit lower triangular matrix, the multi-sensor measurement model with correlated measurement noises is transformed to an equivalent pseudo one with uncorrelated measurement noises; then based on the Markov estimation, a new multi-sensor data fusion method with correlated measurement noises is proposed. Compared with the Markov estimation data fusion method which uses the measurements of original sensors directly, they have the same computational precision, but the new method reduces the computational complexity greatly. Numerical simulation results are provided to demonstrate the validity of the new method.