Weighted measurement fusion estimation algorithm with correlated noises and its global optimality

In view of the multisensor linear discrete time-invariant stochastic control system with correlated noises and different measurement matrices for every sensor,a new weighted measurement fusion estimation algorithm is presented by using the full-rank decomposition of matrix and the weighted least square theory.The newly presented algorithm firstly converts the measurements of many sensors into an equivalent sensor,which is then estimated.The estimating result is proved to be equivalent to the centralized fusion steady-state Kalman estimating result,so that it also has the asymptotic global optimality.It can obviously reduce the computational burden,so it is convenient for application in real time.A simulation result shows the effectiveness of the proposed algorithm.