Abstract:

In response to the problem that traditional grey theory based track association algorithms do not make full use of track history information and reduce association accuracy in scenarios such as dense targets and time-varying noise covariance, a dynamic estimation feedback mechanism based on traditional grey analysis methods is added. The improved algorithm introduces Sage-Husa estimator to estimate real-time sensor noise covariance as a basis for evaluating output data quality, and uses Critical weighting method to convert real-time noise covariance into sequential weights at each time, ensuring maximum utilization of track history information. Simulation has shown that in special correlation scenarios such as dense targets and time-varying noise covariance, the proposed algorithm is significantly superior to traditional grey analysis methods, as well as classical algorithms such as fuzzy and weighted methods, fully demonstrating the performance superiority and robustness of the proposed algorithm.

Key words: track association, multiple sensors, data fusion, grey theory