关键词: 人本感知网络, 隐私保护, 语义攻击, 隐匿空间技术, BP神经网络

Abstract:

As a new type of wireless sensor network system, people-centric sensing network plays a key role in many areas. However, location privacy information is easily leaked due to the semantics-aware threat. Therefore, designing a privacy-reserving scheme which is suitable for people-centric sensing network and can resist the semantics-aware threat is an important and challenging work. Firstly, this paper combines the semantics attributes of geographic position and the obfuscation space technology to reduce the probability of an adversary using the semantics-aware threat to get the user's position privacy information. Then, a new obfuscation space selection algorithm is proposed by using BP neural network technology, which can obtain the appropriate obfuscation space in case of user not transmitting its own actual position information, in order to prevent an adversary to acquire the user’s location privacy information by using the non trusted third party such as obfuscation space generator and server. The theoretical analysis and experimental results show that the scheme can protect the user's location privacy information in a reasonable amount of time and space consumption.

Key words: people-centric sensing network, privacy protection, semantics-aware threat, obfuscation space technology, BP neural network