An Improved Multisensor Data Fusion Method
Xin Yang · Modern Radar · 2007
The optimal weighting algorithm needs to know the variance of noise,and its weight is unchanged during the course of data fusion.FKF(federated Kalman filter) algorithm requires knowing the covariance of the noise.The performance of the two algorithms would decline greatly if the difference between the priori information of noise and the actual value were large,and the noise of sensor system were correlated.In order to overcome these above shortcomings,a novel mutisensor data fusion method is proposed in this paper,which combines the dynamic optimal weighting and least-square filter techniques.The weight used in this method can be adjusted according to the envelope fluctuation of the sensor signal.The statistics information of the noise is not required.The new algorithm performs well and is robust while the noise is correlated.Simulation results indicate its effectiveness.