计算机与通信技术

Multi-sensors fusion performance Bayesian network assessment of airborne based fusion model

对贝叶斯网融合模型的机载多传感器性能评估进行了研究。首先对机载传感器的特性进行了分析，利用专家知识进行网络建模。给出了一种利用本地混淆矩阵（LCM）和髟次测量进行计算得出的全局混淆矩阵（GCM）计算方法，随后又给出了一种利用K次GCM和LCM迭代计算K＋1次GCM的计算方法。给出一些性能指标并利用迭代计算方法对模型分别进行了目标属性未知：目标属性为我、目标属性为敌三种场景下各传感器性能计算，对结果进行了分析并和真实值进行了比较分析，证明了结果的准确性。

Multi-sensors fusion performance assessment of airborne based Bayesian network fusion models was re- searched. Firstly, the net model was built with expert knowledge after the behaviors of airborne sensors were ana- lyzed. Then, the method for calculating the global confusion matrix （GCM）was given by the calculation with the local confusion matrix（LCM） and k measurements ,followed by the method for calculating the GCM of k ＋ 1 with the i iter- ative caculation of LCM and GCM. Some performance indicators were given, and the iterative calculations for the sensors performance under such three scenes, target is unknown, target is ally, target is enemy, were conducted. The results were comparatively analyzed with true values, and the accuracy of the results were verified.