大规模在线监控设备的搜索与识别

In the cyber space, search and exploitation of large-scale online resources are giving the guarantee for ensure national security of critical infrastructure. Surveillance devices that have been the most important and pervasive physical devices, bridging between the physical world and the cyberspace, are widely used in various domains, including government agencies, transportation, plant (Water & Electric), schools and families. Surveillance device discovery, as the first barrier protection for the infrastructure, is becoming very important to national security and social development. There are several challenges and problems in search and recognition of large-scale surveillance devices in the cyber space. We propose a scheduling optimization algorithm for device discovery based on active scanning and passive listening to improve the effectiveness of the device search and reduce the time latency. We propose the algorithm of nonparametric Bayesian based on text-based techniques to discover the unknown surveillance devices in the cyber space. We propose a surveillance device fingerprint model based on deep learning algorithms to improve the accuracy of recognition of a wide variety of surveillance devices. We build a prototype system to verify the relevant technical equipment and search program. We expect to publish 6-8 papers on international conferences and important journals, propose more than 2-4 national patents, and make efforts to produce a certain influence in related research fields.

大规模网络空间在线资源的搜索与利用，是保护国家关键基础设施安全的第一道屏障。在线监控设备连接着物理世界和网络空间，广泛应用于各个领域，包括政府机关、交通、工厂（水利电力）、学校和家庭，已经成为数目众多、影响力最大的物理设备。监控设备的搜索与识别是设备保护的前提，对国家安全、社会发展有着极为重要的意义。针对大规模在线监控设备的搜索与识别所面临的问题与挑战，本课题研究以下内容：研究基于主动发现和被动侦听的监控设备搜索策略调度优化算法，提高设备搜索的实时性和有效性；研究基于文本分析的非参数贝叶斯算法，发现网络空间中的未知监控设备；研究基于深度学习的监控设备指纹模型，提高识别种类繁多的监控设备的准确率；并搭建原型系统验证，验证设备搜索的相关技术方案。本课题预期在国内外重要刊物和会议上发表6-8篇学术论文，申请2-4项国家专利，力争在国际相关研究领域产生一定的影响力。

大规模网络空间在线资源的搜索与利用，是保护国家关键基础设施安全的第一道屏障。在线监控设备连接着物理世界和网络空间，广泛应用于各个领域，包括政府机关、交通、工厂（水利电力）、学校和家庭，已经成为数目众多、影响力最大的物理设备。监控设备的搜索与识别是设备保护的前提，对国家安全、社会发展有着极为重要的意义。针对大规模在线监控设备的搜索与识别所面临的问题与挑战，本课题研究以下内容：高效与实时性的大规模在线监控设备识别模型；研究未知监控设备的识别模型和算法；研究精细化识别在线监控设备的指纹模型，提高识别种类繁多的监控设备的准确率；并搭建原型系统验证，验证设备搜索的相关技术方案。本课题预期在国内外重要刊物和会议上发表10篇学术论文，申请2项国家专利。