Track Association and Fusion Based on Information Demand Analysis

The paper presents a track association and fusion algorithm based on information demand analysis in the multi-sensor and multi-target environment. Track association first is done to judge whether the tracks from the different sensors derive from the same target by the method of the nearest neighborhood. Then the algorithm discards the tracks with poor quality by analyzing information demand for the fusion and evaluating the quality of each track using the method of statistical variance analysis. Last, Kalman filter and Simple Fusion strategy are used for the state estimation fusion. Experiment results show the algorithm improves the precision of the final track during the process of track fusion.