贝叶斯网络用于态势估计时,系统参数不能及时调整,无法对未来时刻进行预测。为解决这一问题,对动态贝叶斯网络在战场态势评估中的应用进行了研究,引入时间因素,建立了网络模型,分析了概率参数与推理的过程,并利用卡尔曼滤波模型法对推理进行仿真实验。实验结果表明了动态模型推理的有效性。动态贝叶斯网络可以有效利用侦察数据中的时间信息,实时动态地处理影响分析和决策的种种因素,对指挥员的作战决策具有极大的参考价值。

The Bayesian Network(BN) is not able to predict the future when it is used for situation assessment since the system parameters can not be updated in real-time. To solve this problem, we studied the application of Dynamic Bayesian Network (DBN) in battlefield situation assessment. The factor of time was introduced to construct the network model. The probability parameters and inference procedure were analyzed, and Kalman filter algorithm was used in simulation of the inference. The simulation result proved the feasibility of the dynamic inference model. Since information of time from reconnaissance data can be used effectively in DBN for dynamically processing the factors that have effect on analysis and decision-making, thus it is of great significance for decision-making of commanders.