本文已被:浏览 3725次 下载 2846次
投稿时间:2018-06-19 修订日期:2018-07-05
投稿时间:2018-06-19 修订日期:2018-07-05
中文摘要: 航空运输业在世界经济和社会发展中发挥着重要的推动作用。国际民用航空组织统计显示全球空中交通量大约每15年增加一倍,因此,现有空中交通航行系统的运行能力已经接近饱和。现有的空管自动化系统能自动获取和处理空管信息,经过监视数据融合、飞行数据处理、气象及航行情报处理、告警处理等形成空管综合态势,以供管制员指挥使用,但其仅具备有限的决策支持能力,空管智能化程度较低,不能较好地适应未来空管发展的需要。为了适应未来航空业的快速发展,解决空中交通安全、空域拥挤、航班延误等问题,各国致力于开展解决未来空中交通问题的新技术研究。欧盟在2004年提出“单一欧洲天空空中交通管理研究”,拟重新规划欧洲空域以满足空中交通需求,提高空管系统效能,该研究的关键技术涵盖空中交通管理的4个关键领域:高效的机场运行、高级空中交通服务、优化的空中交通网络服务和可靠的空管基础设施。美国在2005年提出了“下一代航空运输系统”,其核心技术包括广播式自动相关监视、数据通信、航路自动化系统现代化、终端自动化系统现代化更新、NAS语音系统和系统广域信息管理。 2012年,ICAO推出了航空系统组块升级(ASBU)计划,以运行改进为核心,以现有空管技术和新技术应用为手段,提出了机场运行、全球互用的系统和数据、最佳容量和灵活飞行、高效的飞行轨迹4个性能提升领域,每个领域由多条提升路径组成,并根据实现阶段分布在4个组块中,构成52个模块。中国民航也正在实施或规划大量ASBU中的内容以应对中国航空的快速发展。但是,现有空管系统及其未来规划重点在基础设施建设,未涉及太多智能化应用。近年来,在深度学习、高性能计算和大数据的支撑下,人工智能技术在各行业得到快速应用,计算机视觉、语音识别、自然语言处理等技术取得突破性进展并迅速产业化。人工智能技术可以促进空管一些关键技术的发展,提升安全水平,提高管制工作效率,降低管制员工作负荷。以深度学习为代表的人工智能强调在大量先验知识的基础上做出判断和决策,与空中交通管理运行决策过程相契合,人工智能的快速发展产生的巨大效益使智能化空管成为一种必然。因此,在空管技术及人工智能技术基础上,作者提出智能化空管的概念,设计了智能化空管系统总体框架。智能化空管系统总体框架包括感知层、网络层、平台层、应用层和可视化层。感知层的各类通信、导航、监视、气象、无线、图像采集、射频识别等设施设备为空管运行提供基础设施保障。网络层采用专线网络、卫星通信网、新兴的互联网、移动网络等传输信息。智能化空管平台层采用广域信息管理,云计算、智能化大数据挖掘等技术实现信息存储、共享、挖掘等。应用层研究人工智能化技术在管制指挥、空域管理、空中交通流量管理、飞行服务、通航、无人机空管保障中的应用。可视化层通过空管门户、虚拟化可视、空管智能化UI、移动空管应用等方式提供高效快捷的智能化交互。围绕智能化空管概念和总体框架,需要重点研究智能化空管数据处理、智能化辅助决策、空管语音识别和空管指挥机器人等关键技术。智能化空管数据处理研究各类空管数据的获取、处理、传输、交互、智能化挖掘等技术;智能化辅助决策研究智能化冲突管理、智能化空中交通流量管理、智能化规划管理、智能化进离场排序、智能化机场运行等智能化空管应用;针对地空通话在空中交通管制中的重要作用,研究空管语音识别在自动应答机长、空管指挥安全监控中的应用;利用人工智能技术使机器具备空管的感知、规划、推理、行动等能力进行空中交通管理。智能化空管系统总体框架和关键技术为智能化空管系统开发提供理论基础与技术支撑。
Abstract:Air transport plays an important role in promoting the world economy and social development. The statistics of international civil aviation organization show that global air traffic is approximately doubled every fifteen years, and the existing air traffic navigation system is close to saturation. The existing air traffic management (ATM) systems automatically acquire and process air traffic control information for air traffic controller, through surveillance data fusion, flight data processing, meteorology and aeronautical information processing and safety nets processing. However, the limited decision-making support ability leads to the low intelligent degree of ATM system, which can't meet the need of the future development of ATM. In order to adapt to the rapid development of aviation industry in the future, some countries and organizations are committed to carrying out new technologies to solve the problems of air traffic safety, airspace congestion and flight delays. In 2004, the EU proposed the "Single European Sky ATM Research(SESAR)" program and proposed tore-plan European airspace to meet air traffic demand and improve the efficiency of ATM system. The key technologies of the program include four key areas of ATM:efficient airport operations, advanced air traffic services, optimized air traffic network services and reliable ATC infrastructure. In 2005, the United States proposed the next generation air transportation system(NEXTGEN), including ADS-B, data communication, en route automation modernization, terminal automation modernization and replacement, NAS voice system and system wide information management. In 2012, ICAO launched the aviation system block upgrade plan (ASBU). The ASBU involved four aviation performance improvement areas, including airport operations, globally interoperable systems and data, optimum capacity and flexible flights and efficient flight path. Each area consists of multiple threads and distributed in four blocks according to the implementation stage. CAAC is also implementing or planning a large number of ASBU modules to cope with the rapid development of Chinese civil aviation. The existing ATM system and its future planning mainly focus on the infrastructure construction, without enough intelligent applications. In recent years, with the support of deep learning, high performance computing and big data, artificial intelligence technology has been rapidly used in various fields. Especially, technologies of computer vision, speech recognition and natural language processing have made breakthrough and rapid industrialization. Artificial intelligence, represented by deep learning, emphasizes that judgment and decision-making based on a large number of prior knowledge, which is consistent with the decision-making process of ATM. Therefore, artificial intelligence technology can promote the development of air traffic control key technologies, improve safety and the efficiency of air traffic control and reduce the workload of air traffic controller. The great benefit of the rapid development of artificial intelligence makes the intellectualized ATM a necessity. In this paper, the concept of intellectualized ATM and the overall framework of intellectualized ATM system were presented. The overall framework of intellectualized ATM includes the perception layer, network layer, platform layer, application layer and visual layer. All kinds of communication, navigation, surveillance, weather, wireless, video capture, radio frequency identification and other facilities of the perception layer offer the infrastructures for ATM. The network layer transmits information by using special line network, satellite communication network, Internet, mobile network and so on. The platform layer achieves information storage, sharing and mining by using of SWIM, cloud computing, intelligent big data mining. Application layer studies the application of artificial intelligence technology in air traffic control, airspace management, air traffic flow management, flight service, general aviation and unmanned aerial vehicle. The visual layer provides efficient and intelligent interaction through the portal, virtual visualization, intelligent UI and mobile applications. The research direction of intellectualized ATM includes intelligent ATM data processing, intelligent decision-making, air traffic control speech recognition and air traffic control robot. The intelligent ATM data processing involves various kinds of ATM data acquisition, processing, transmission, interaction and intelligent mining. Intelligent decision-making focuses on intelligent conflict management, intelligent air traffic flow management, intelligent planning and management, intelligent AMAN and DMAN, intelligent airport operation, etc. In view of the important role of ground-to-air communication in air traffic control, the intelligent simulation pilot and air traffic control safety monitoring based on automatic speech recognition should be studied. The intellectualized ATM system has the capabilities of air traffic control perception, planning, reasoning and action based on artificial intelligence. Thus, the overall framework and key technologies provide the theoretical basis and technical support for the development of intellectualized ATM system.
keywords: air traffic management artificial intelligence deep learning reinforcement learning ATM robot
文章编号:201800688 中图分类号: 文献标志码:
基金项目:国家空管委科研资助项目(GKG201403004);国家重大科学仪器设备开发专项资助(2013YQ490879)
作者简介:杨红雨(1967-),女,教授,博士.研究方向:空管信息智能处理技术;视觉合成.E-mail:yanghongyu@scu.edu.cn
引用文本:
杨红雨,杨波,武喜萍,余静.智能化空管技术研究与展望[J].工程科学与技术,2018,50(4):12-21.
YANG Hongyu,YANG Bo,WU Xiping,YU Jing.Research and Prospect of Intellectualized Air Traffic Management Technology[J].Advanced Engineering Sciences,2018,50(4):12-21.
引用文本:
杨红雨,杨波,武喜萍,余静.智能化空管技术研究与展望[J].工程科学与技术,2018,50(4):12-21.
YANG Hongyu,YANG Bo,WU Xiping,YU Jing.Research and Prospect of Intellectualized Air Traffic Management Technology[J].Advanced Engineering Sciences,2018,50(4):12-21.