离散模糊动态贝叶斯网络用于无人作战飞机目标识别

为了提高无人作战飞机的自动目标识别能力,文中提出采用离散模糊动态贝叶斯网络,对若干可观测的目标特征参数进行综合推理出目标的类型。推导了离散模糊动态贝叶斯网络的推理算法。建立了目标识别的离散模糊动态贝叶斯网络模型。仿真结果表明,该方法得出的推理结果与理论分析完全一致,而且能够将各种并不显著的目标特征进行综合,使得各种特征及不同时刻的同一特征互相修正补充,克服了依靠单一特征进行目标识别的局限。

Improving Identification Capability of UCAV (Unmanned Combating Air Vehicle)

Existing identification capability of UCAV,as occasionally found in the open literature,is,in our opinion,not satisfactory and can be much improved.We present a method that we think does much improve such capability.Our new method uses the discrete fuzzy dynamic Bayesian network as the tool of inference to determine the category of the target from just several observable parameters that characterize the target.Simulation results show preliminarily that:(1) our method can integrate various rather obscure target characteristics and make them conspicuous as a whole;(2) various characteristics reinforce and modify each other;(3) the values of the same characteristic at different instants reinforce and modify each other;(4) the above three achievements overcome the limitation unavoidable if identification relies on only a single characteristic(many existing identification methods just rely on a single characteristic).