Research On Privacy Protection In Wireless Sensor Networks

Wireless sensor networks, commonly abbreviated as WSNs, are usually consisted of numerous self-organized low-cost and feature-rich sensor nodes. In the recent few years, WSNs have tremendous promising alternative applications and can be widely deployed in many application areas, such as military battlefield, health cares, environmental protection, smart home, and so forth. Therefore, WSNs have drawn board attention from both industry and academia. With the rapid development and proliferation of WSNs, the increasing serious privacy concerns have gradually restricted the deployment and application of WSN. However, compared with traditional networks, WSNs are resource-constrained and application specific, which determines that their privacy problems were significantly distinguishable and unique, leading it more difficult to effectively apply existing privacy protection mechanisms and algorithms to address related problems. Consequently, it brings emergent requirements and great challenges for designing privacy protection solutions within WSNs.In this dissertation, we focus on the privacy protection concerns in WSNs, and carried out our in-depth study aiming at addressing several key technical issues. We make our great effort to analyze the privacy threat and its corresponding protection requirements in WSNs, and then further conducted location privacy preserving inference control mechanisms towards successive source locations, designed conditional identity privacy preserving protocols for vehicular sensor networks, and proposed task allocation and scheduling approaches for data privacy protection in multimedia sensor networks. Towards these issues, based on well-established mathematical models, we proposed corresponding solutions, algorithms and protocols. Our contributions can be summarized as follows:According to different protection objects, the privacy problem in WSNs can be classified into three categories:data privacy, location privacy and identity privacy. In this thesis, the privacy protection concerns in WSNs are analyzed in a comprehensive way, and followed by classification, analysis and reviews on typical schemes and mechanisms. We hope that our work may be helpful for domestic researchers on this hot topic.Location Privacy is one of the major challenges in surveillance sensor networks. Currently, most research efforts focus on protecting current location, and ignore the internal relationship among the successive locations. To date, given a sequence of past observations, abundant techniques are available to infer current or future locations of an object, which may lead possible infer attacks and bring in serious privacy and security concerns. In this thesis, we for the first time proposed the successive location privacy infer attack and protection problems, and built the K-successive privacy model based on in-depth analysis of these problems. In this model, we observed that there is an intrinsic trade-off between the number of data to be published to the interested parties and the privacy preservation of the object. We then formulated the maximum publishable sequence(MPS) problem, and show it is NP-complete. Thereafter, we analyzed the online MPS problems, and developed optimal solutions to p-relation cases through dynamic programming and two heuristics for the general case. We then designed the P-preserving protocol for distributed privacy aware data collections. Extensive simulations and analysis demonstrate the effectiveness of our proposed algorithms and protocols.Identity privacy concerns are very critical and fundamental aspects in emerging vehicular sensor networks. So far, existing privacy preservation schemes are unable to be deployed in common urban scenarios, and therefore their application becomes severely constrained. By analyzing the inherent requirements for deploying privacy preservation schemes in common urban scenarios, we propose SCPP, a novel short-time certificate privacy preservation protocol. SCPP is not only well suitable and easy-to-deploy in common urban scenarios, but also satisfy the requirements of conditional privacy preservation and identity revocation as well. Extensive analysis and comparison results show that, compared with the current best available privacy preserving protocol, our SCPP protocol performs better in terms of both computational and communication cost.With the rapid development of multimedia sensor networks, its data privacy protection problems bring great challenges to task allocation and scheduling. To this end, by analyzing its real-time and low-power requirements, the task allocation and scheduling problem for data privacy protection in multi-hop multimedia sensor networks was found to be topology aware. By investigating multi-objective optimization and constraint modeling techniques, we formally model TATAS, the Topology-Aware Task Allocation and Scheduling problem, and prove it is NP-complete. Then, an efficient three-phase heuristic solution, named TATAS-3H, was proposed to solve the TATAS problem. Experimental results show that, as compared with traditional approaches, our technique can achieve significant energy saving and effectively meet the real-time requirements of data privacy protection as well.