面向智能辅助决策的飞行器安控系统构建研究

Construction of Safety-control System for Aircraft Based on Intelligent Decision Supporting

新一代人工智能技术能够提高飞行器试验安控系统的辅助决策能力,满足地面安控目标数量增多、故判规则量增大、故判时效性高等新需求,具有广阔的应用前景。在对飞行器试验地面安控需求分析的基础上,设计面向智能辅助决策的地面安控系统总体框架,增加了数据与模型管理模块;接着构建了辅助决策知识库,主要包括故障弹判据、速度判据、位置判据和预示落点判据等规则库;在此基础上提出数据和规则协同驱动的故障预测方法。该系统既利用历史积累试验数据对深度神经网络模型进行训练,又利用规则库提高专家经验对模型预测精度的贡献,能够提高故障智能预测能力,满足飞行器安控智能辅助决策需求。

The decision supporting capability of aircraft test safety-control system can be enhanced by using new-generation artificial intelligence technologies,which can also meet the needs of increasing in ground safety-control quantity,rule numbers en-larging and higher timeliness of fault diagnosis.So they bring extensive application prospect.The architecture of ground aircraft safe-ty-control system aiming at intelligent decision supporting is designed based on the new requirements analysis,which adds the data and model management module.Then the decision supporting base is constructed which contains the fault diagnosis rules of aircraft itself,velocity,location and impact point prediction feature recognition.A new fault prediction method is proposed based on the co-operation of rule based and data-driven.The intelligent ground aircraft safety-control system can use the historical test data to train the deep neural network and also use the rule base to enhance the expert knowledge contribution for the model prediction accuracy,which can enhance the fault intelligent prediction ability and meet the intelligent assistant requirements of aircraft safety-control.