Dynamic Track Fusion Algorithm Based on Information Quality Selection

Abstract

The traditional track fusion algorithm does not fully consider the situation that the accuracy of sensors and the measurement loss lead to the track quality degradation. In order to improve the performance of dynamic tracking, a dynamic track fusion algorithm based on information quality selection is proposed. The algorithm obtains the local track and information entropy by interacting multiple model compensation filtering, and then uses the information entropy to measure the quality of the local track. The local track with good quality is selected according to the double threshold. Then, the information entropy normalization result is assigned to the weight of the sensor to realize the dynamic matching of the weight. The simulation results show that the algorithm outperforms the known track fusion algorithm in tracking maneuvering targets with different sensor accuracy and measurement losses.

Keywords

• |multi sensor|measurement loss|information entropy|track selection|track fusions|target tracking