Multi-sensor centralized fusion tracking with correlated measurement noises

In practical multi-sensor fusion target tracking, the measurement noises of different sensors are always correlated. By using the Cholesky factorization and inverse method for unit lower triangular matrix, the multi-sensor measurements with correlated measurement noises are transformed to equivalent pseudo ones with uncorrelated measurement noises; then based on the Kalman filtering, a new multi-sensor fusion target tracking algorithm with correlated measurement noises is proposed. Compared with the existing centralized fusion algorithm and the centralized fusion algorithm which uses the measurements of original sensors directly, they are equivalent in computational precision, but the new one reduces the computational complexity greatly. Numerical simulation results are provided to demonstrate the validity of the new algorithm further.