SELF-TUNING WEIGHTED MEASUREMENT FUSION WHITE NOISE DECONVOLUTION ESTIMATOR

For the multi-sensor linear discrete time-invariant stochastic systems with correlated measurement noises and unknown noise statistics, an on-line noise statistics estimator is obtained using the correlation method. Substituting it into the optimal weighted fusion steady-state white noise deconvolution estimator based on the Kalman filtering, a self-tuning weighted measurement fusion white noise deconvolution estimator is presented. By the Dynamic Error System Analysis (DESA) method, it proved that the self-tuning fusion white noise deconvolution estimator converges to the steady-state optimal fusion white noise deconvolution estimator in a realization. Therefore, it has the asymptotically global optimality. A simulation example for the tracking system with 3 sensors and the Bernoulli-Gaussian input white noise shows its effectiveness.