基于弹性变结构DDBN网络的空战目标识别

采用离散动态贝叶斯网络的直接推理方法作为基础,提出弹性变结构离散动态贝叶斯网络的概念,构建了空战目标识别的弹性变结构离散动态贝叶斯网络模型,给出了相应的推理算法,以此克服了离散静态贝叶斯网络和定结构离散动态贝叶斯网络在目标分类识别过程中出现的问题。通过仿真结果对比,表明该方法可以综合各个时刻各种被观测到的确定和不确定信息,从而更为有效的实现目标的分类和识别。

Target Recognition in Air to Air Combat Based on Elastic Variable Structure Discrete Dynamic Bayesian Networks

Based on the direct reference algorithm of Discrete Dynamic Bayesian Networks (DDBN), the concept of elastic variable structure DDBN was proposed and the corresponding reference algorithm was studied. Using the elastic variable structure DDBN, the target identification model was established and it was capable to overcome the problems perplexed by Discrete Static Bayesian Networks and constant structure DDBN. Simulation results show that this method can synthesize various deterministic and non-deterministic information in every time period so as to make the target classification and identification more effective.