Multi-sensor distributed information fusion unscented particle filter

In this paper, an unscented particles filter based distributed information fusion is proposed for state estimation problem of nonlinear and non-Gaussian systems. It uses unscented Kalman filter algorithm to update particle; then calculates local state estimated values by particle filter. The system fusion estimation is obtained by applying the fusion rule weighted by scales. The simulation results show that compared with single sensor, the proposed algorithm improves the accuracy of filter.