Data Fusion Simulation Algorithm Based on Confusion Matrix Performance Metrics

摘要

Data fusion is an increasingly important capability in operational command andcontrol systems. A fundamental function of data fusion processing associates unidentified contact reports from multiple sensors to a common target, so that the commander has a consistent tactical picture from which to make decisions. Clearly computer simulations should model the data fusion aspect of the command and control process. However, since data fusion algorithms are typically computationally intensive, they have been neglected in these simulations. This paper describes a new approach for low resolution simulation of data fusion processing. Operational data fusion systems are evaluated on known data sets using four performance metrics calculated from the confusion matrix-a tabulation of the number of contact reports from each ground truth track assigned to each constructed track. The simulation algorithm described in this paper makes probabilistic data fusion decisions based on values input for these metrics. This paper also describes an object-oriented implementation of the algorithm in C++ and presents numerical results from a simulation study of the algorithm itself, showing its ability to produce data fusion results with the input performance characteristics. (KAR) P. 1.