Research On Secure Localization In Wireless Sensor Networks

Sensor node localization is a fundamental technique in WSN. However, unsupervised sensor nodes are susceptible to the from malicious nodes and captured nodes due to its inherent openness. The invalid and erroneous node localization can have a destructive impact to WSN application, which can lead to false supervision results, crack down of network etc. Therefore, How to provide a secure localization technique in resource limited WSN is a challenging and valuable work.Current secure localization algorithms have solved this problem in different ways, but all have the following disadvantages: (1) dependent on the deployment policy. If the deployment model can not simulate the real deployment, it can lead to localization error, such as CSB, SLA. (2) Requirement of special time synchronization equipment. Since it has strong dependent on hardware, it does not suitable for the wide deployment of WSN. (3) It does not work well fro colluded attack. Normally, localization error becomes higher as the ratio of colluded nodes get higher. For those whose ratio exceed to 50%, algorithm does't work totally. Therefore, realization of a algorithm which can defense those attacks without special equipment and reduce cost as much as possible is a major challenge. To solve the above problem, this dissertation proposed a secure localization algorithm ETMCA based on redundant information of neighbors of non-beacon nodes and the consistence of data. ETMCA can defense different types of independent attacks and colluded attacks without special equipment.ETMCA has two phases: information set clustering and secure validation. Information set clustering based on data consistence. When it detects independent attack, it can use the unique cluster to locate itself. Otherwise, when it detects colluded attacks, it goes into the secure validation phase to further process. In secure validation phrase, ETMCA use the redundant information of located none-beacon nodes to find the secure localization information and calculate the localization results. Therefore it solves the problem. Simulation results show: when localization error subject to uniform distribution -ξmax < <ξmax and equals 0.818. (1) For independent attacks ETMCA has a similar performance with current existing algorithms. (2) For the colluded attack whose malicious beacon node ratio exceeds 50%, ETMCA's detection rate can reach 90%, localization error is in a acceptable range. Meanwhile, other algorithm does't work in this case. Designing a clustering algorithm which can balance energy consumption and localization accuracy so that algorithm complexity can be largely reduced. In this way, we can detect most of the colluded attacks and prolong the network's life time.