面向复杂低空密集飞行的航空器自主安全间隔保持理论与方法研究

The increasingly diversified and large-scale characteristics of flight activities in the complex low-altitude airspace environment have greatly increased the risk of flight collision, posing severe new technical challenges to the operation safety of low-altitude airspace. Aircraft flight safety separation maintenance for intensive flight in low-altitude airspace has attracted much attention from international air traffic control academia and industry community in recent years.Because intensive flying in the low-altitude airspace present a highly complex and highly dynamic space-time mutual restriction feature, the uncertainty of the airspace security evolution situation and heterogeneous aircraft flight behavior diversity are intertwined. The traditional aircraft flight conflict detection and resolution method can not effectively deal with the potential systemic threat of intensive flying in complex low-altitude airspace.The project is to build a low-altitude intensive flight safety situation evolution model and evaluation index system, analyze the collision risk key elements and their interaction mechanism, reveal the airspace safety risk dynamics evolution characteristics and mining heuristic knowledge for flight collision risk prediction and safe separation keep.Furthermore, according to the characteristics of low-altitude operation, the judgment basis of aircraft flight conflict prediction will be designed based on the statistical classification theory of machine learning, and the flight collision risk classification prediction method is studied according to the different anti-risk capabilities of heterogeneous aircraft.Finally, considering the aircraft safety risk level and efficiency target comprehensively, we will study the multi-agent autonomous safety separation maintenance method based on satisfactory game and negotiation. The research of this project will lay a solid theoretical foundation for the security control of low-altitude airspace operation.

复杂低空空域环境下飞行活动多样化、规模化特点日益凸显，极大增加飞行碰撞风险，对低空空域运行安全提出了严峻技术挑战。面向低空密集飞行的航空器飞行安全间隔保持技术近年备受国际空管学术界和工业界关注。由于低空空域密集飞行呈现出高度复杂、高动态的时空相互制约性，空域安全态势演化的不确定性与异质航空器飞行行为的多样性共同作用，传统方法难以有效应对潜在系统性威胁。本项目构建低空密集飞行安全态势演化模型与评估指标体系，分析复杂受限低空下导致冲突风险关键要素及相互作用机理，揭示空域安全风险态势演化特性与规律，挖掘启发式知识指导方法设计；基于机器学习统计分类理论设计航空器飞行冲突预测判定依据，针对异质航空器不同抗风险能力研究飞行碰撞风险分级预测方法；最后综合考虑航空器安全风险等级与效率目标，研究复杂低空密集飞行下基于满意博弈与协商的多主体自主安全间隔保持方法。本项目将为低空运行安全管控奠定坚实的理论基础。

复杂低空空域环境下运行场景日益丰富、航空器多样、飞行活动规模化特点日益凸显，极大增加飞行碰撞风险，对低空空域运行安全提出了严峻技术挑战。面向低空密集飞行的航空器飞行安全间隔保持技术近年备受国际空管学术界和工业界关注。由于低空空域密集飞行呈现出高度复杂、高动态的时空相互制约性，空域安全态势演化的不确定性与异质航空器飞行行为的多样性共同作用，传统基于人决策的空中交通管理和低空服务保障等方法难以有效应对潜在系统性威胁。. 围绕通用航空及民用无人驾驶航空器发展需求，面向密集飞行的低空飞行碰撞风险预测与冲突解脱理论和方法进行系统的国内外调研与综述，结合低空试点空域、应急救援、通用机场、城市物流等典型运行场景，收集、分析我国低空空域运行的相关历史数据，构建低空密集飞行空域结构模型，分析复杂受限低空下导致冲突风险关键要素及相互作用机理，挖掘启发式知识指导方法设计。针对非合作航空器研究基于射频信号倒频谱及模糊评价模型的无人机监测评估方法，针对合作航空器基于机器学习提出了异质航空器飞行冲突预测方法。最后综合考虑航空器安全风险等级与效率目标，构建战略战术多层冲突探测与解脱方法体系，从静态空域结构提出分层空域模型和优化设计方法，针对复杂密集飞行提出了基于混合人工势场与蚁群算法的多航空器冲突解脱方法，进一步面向应急救援场景提出基于Agent的低空实时飞行冲突探测与解脱方法，最后搭建了复杂低空多航空器避撞与验证平台针对场景和方法开展测试验证。. 截至2022年底，依托本项目研究，形成多项学术成果，发表学术论文18篇，其中SCI检索论文5篇，EI检索论文9篇，中文核心检索论文4篇；授权国家发明专利1项，申请国家发明专利6项，授权软件著作权4项；获得中国航空学会、中国航空运输协会科技进步奖二等奖共计2项，起草形成中国民用航空局行业标准2项，参与形成行业规范性文件1项。本项目研究为低空运行安全管理完善理论基础，为低空空域深化改革提供技术支撑，促进空中交通管理的理论与方法创新，在低空航线划设与空管运行方面为行业监管提供了行业标准和规范。