Abstract: Using traditional localization algorithms the location error is high for moving target. To solve this problem, an improved weighted centroid localization algorithm based on Received Signal Strength Intensity (RSSI) is proposed, and an improved Unscented Kalman Filter (UFK) algorithm is combined for RSSI filtering. For the defect that traditional centroid localization algorithm can only statically set the weights, in this improved algorithm, the localization error is used to correct the distance, which can effectively improves the localization accuracy. For the standard UKF algorithm, the minimum skewness strategy is used in the sampling process, and the adjustment factors are introduced to ensure the positive semidefinite of the prediction covariance matrix; simultaneously, the fading memory filtering method is used in filter updating process, which effectively inhibits the filter divergence and improves the filtering accuracy. Simulation results prove the correctness and effectiveness of the proposed algorithm.

Key words: information processing, moving target localization, weighted centroid algorithm, unscented Kalman filter(UKF)