动态贝叶斯网络的无人机炮火校射

动态贝叶斯网络(Dynamic Bayesian Networks-DBNs),是对具有随机过程性质的不确定性问题进行建模和处理的一个有力工具.提出将隐马尔可夫模型(Hidden Markov Models-HMMs)图形模式与贝叶斯网络结合起来构成DBN,将其用于无人机照相侦察情报的推理分析,决定炮火优先打击区域.首先建立动态贝叶斯网络的战场态势变化模型,而后应用HMM的推理算法获得当前隐含序列最优估计,且可预测出未来战场态势.最后应用模糊推理获得优先打击的区域号.仿真结果表明了模型的可行性.该方法有效解决了贝叶斯网络对于瞬间变化战场态势推理的不足的缺陷,为炮兵指挥员更好地运用火力,分出主次奠定了基础.

Adjust Fire Goal with UAV based on Dynamic Bayesian Network

Dynamic Bayesian Networks are a powerful methodology for representing and computing with uncertain problem of stochastic processes.We combine Hidden Markov Models with Bayesian Network to rearch photo gained by UAV,and build a battlefield dynamic model base on Dynamic Bayesian Networks.In the end,we use a Viterbi algorithm to inference and get the best estimate about hidden sequence.The next step,we can use Fuzzy inference to get number of area that should shoot first.The model can provide correct and useful information for commander so that he can separate important things from general and decide how to use shellfire better.