Research On Accurate Dynamic Localization And Security In Wireless Sensor Networks

As the rapid development of information technology, wireless communications, integrated circuits and MEMS (micro-electro-mechanism system) technology has made tremendous progress, designing low-cost, low power, small size and multifunctional micro sensor nodes has become possible. The sensor nodes have capabilities of information perception, data processing and wireless communications. In addition, the research progress of distributed information processing and self-organizing networks technology has provide technical support for large-scale wireless nodes cooperative work, data processing and transmission and self-organizing network building. All these technologies have produced wireless sensor network (WSN). As a new way of information acquisition and processing, wireless sensor network has many advantages, such as don't need fixed network infrastructure, rapid deployment and strong anti-destructive, it has broad application prospects on the battlefield surveillance, environmental monitoring, medical aid, science exploration and disaster early warning. Wireless sensor network has the following characteristics, the large number of sensors, broad coverage, dynamic topology, network self-organization, limited hardware resources, data-centric and so on.Localization technology is the basis for other applications in wireless sensor networks, acquisition of nodes'physical location is prerequisite for many high-level applications, from the earliest days of sensor networks research, many scholars starting the node localization algorithm research in different perspectives and applications, propose a number of localization algorithms and localization system. According to wireless sensor nodes whether or not use the distance or angle measurement technique, the localization algorithms are divided into range-based localization algorithms and range-free localization algorithms; according to the location that algorithm executed, localization algorithms can be divided into centralized and distributed localization algorithms; according to different reference coordinate systems, localization algorithms can be divided into absolute position localization algorithms and relative position localization algorithms; according to whether or not use the mobile network nodes, localization algorithms can be divided into static and mobile localization algorithms. The performance of localization algorithms could directly affects its availability, there are a few number of evaluation criteria, such as localization accuracy, localization scale, reference node proportion, node deployment density, fault tolerance, adaptability, power consumption, computational complexity and so on.In research for wireless sensor localization algorithms, DV-Hop and Centroid-based algorithms are fundamental methods in wireless sensor networks. They have their own localization characteristics and corresponding disadvantages. In order to make more effective use of the two algorithms advantages, avoid the disadvantages of localization accuracy and power consumption, on the basis of the two algorithms, a range-free hybrid localization algorithm CDHL (Centroid and DV-Hop Hybrid Localization) has been proposed in this paper. CDHL algorithm synthesizes centroid and DV-Hop algorithm in different stages, its main principles are hierarchical anchors and iterative refinement. For unknown node, it first judge how many anchor nodes are around in the scope of one hop. If the node count is not smaller than a certain integer N, select the centroid algorithm to calculate the unknown node's position. Meanwhile, all the nodes that have been located (nodes that are located whether by centroid algorithm or by DV-Hop localization algorithm) will be upgraded to secondary anchor nodes. These secondary anchor nodes will help calculate the position coordinates of other nodes in next iterative calculation. If the number of anchor nodes around unknown node is less than N, select the improved DV-Hop localization algorithm to calculate the unknown node's position. A new strategy to calculate the"average distance per hop"for an unknown node is proposed. For each unknown node, select the anchor nodes which have the least distance from the unknown node and whose hops to this unknown node emerge most frequently. Then, make use of these preferred anchor nodes to calculate the average distance per hop. For some high frequency anchor nodes, their estimate coordinates are calculated through the CDHL algorithm, use these anchors'estimate coordinates and the actual coordinates to calculate the correction factors, then use these coordinate correction factors to revise the estimated coordinates of unknown nodes. These factors will further improve the localization accuracy. All unknown nodes that have been localized will be seen as the secondary anchor node, then start the next round of unknown nodes iterative calculation, till all unknown nodes have been localized. In this paper, the algorithm average localization accuracy is analyzed, the algorithm energy consumption is discussed, simulation results show that CDHL algorithm has higher localization accuracy and lower energy consumption level in variety of sensor networks deployment situations, it could increase the average localization accuracy by 5% -7% than DV-Hop algorithm.Localization problem is an important research challenge in wireless sensor network, the early localization algorithms are lack of effective energy management and localization accuracy is still insufficient in the practical applications. In recent years, many scholars have proposed a few of new localization algorithms or improvements on traditional methods, and these new methods received good effects. In this paper, for the relations between localization error of unknown node and the distance from neighboring anchor node, a theory that localization error of unknown node is consistency changing with the distance from neighboring anchor node is proposed, and proof of the theory is given. In this theory, if it is assumed that physical coordinates of some anchor nodes are unknown, we could calculate their estimate coordinates by other anchors'physical coordinates through distance estimate method, there will be localization error distance between its physical coordinates and estimate coordinates, then the unknown nodes nearby this anchor will be share this localization error if we calculate the unknown node estimate coordinates by the same localization algorithm, and theoretically the closer to this anchor, the more similar the localization error will be. Based on coordinate error theory above, the methods that anchor nodes multilevel classification depend on different distribution and estimate coordinate correction according to different scope level of anchors are proposed. According to different correction levels, unknown nodes need to calculate different amount of correction reference vectors. However, if the correction vectors are too many or anchors too far from unknown nodes, they may reduce coordinate localization accuracy. So, for a certain unknown node, after anchor nodes multilevel classification depend on different distance scope, only select the appropriate level could improve the localization accuracy. Based on the theory and methods above, MCL (Multilevel Correction Localization) algorithm is proposed in this paper. The algorithm can classify anchor nodes depend on hops from unknown node and select the appropriate level based on node communication radius and the network coverage. The coordinate multilevel correction experiments show that correction vector could improve the localization accuracy. MCL algorithm localization accuracy experiments show that MCL algorithm is more efficient, more accurate and more energy efficient than DV-Hop algorithm. The average localization accuracy of MCL is approximately 7% to 9% higher than DV-Hop. In addition, the multilevel coordinate correction mechanism is also suitable for many other anchor-based localization algorithms.Wireless sensor networks are often carried out in three-dimensional space, so three-dimensional sensor node localization algorithms are more practical. In this paper, the three-dimensional localization technology of wireless sensor network is discussed, the three-dimensional sensor node localization model is described, the three-dimensional space node localization security problems and possible solution strategies are analyzed. The corresponding solution code realization is given. In localization process, a number of node security problems are proposed, such as anchor node failure problem, anchor node issues false localization information problem, malicious node false accusation problem and so on. Anchor node failure is the most common security problem, include several different types. Compared with other nodes, anchor nodes need to transmit more information and face more energy depletion attack. Anchor nodes are malicious captured or destroyed will lead to node failure. The captured anchor nodes may be hardware tampered and issues false localization information. If the localization algorithm includes some simple reliability self-inspection and mutual inspection mechanism, malicious anchor nodes could false accusations other anchor node issues false localization information. A malicious anchor node false accusation other anchor nodes will confuse unknown nodes localization reference selection strategy, will lead to localization error or even cause the failure of the localization algorithm. In order to avoid malicious anchor node localization interference, anchor node security detection is needed to reduce the localization error before unknown node localization estimation calculation. To achieve this objective, a dynamic group filtering anchor node security detection algorithm is proposed in this paper. According to the distance, anchor nodes are grouped to 4-dimensional localization matrix, the matrix that exceed the normal error levels maybe contain malicious anchor node, then vote this matrix, at last most voted node is malicious node. However, the malicious anchor nodes may execute malicious vote, that is to say a malicious anchor nodes calibration of other benign anchor nodes and vote them, the vote results will not be fairness and can't filter out the truly malicious nodes. For this situation, a voting selection strategy is proposed in this paper, under this strategy to vote and vote supervision could ensure that malicious anchor nodes will receive the highest number of votes and will be screened out. Based on the thoughts above, a three-dimensional space DGFSL-3D (Dynamic Group Filtering Security Localization) algorithm is proposed. Simulation experiments show that the anchor node security detection algorithm and voting selection strategy can effectively reject the malicious nodes and significantly reduce the malicious node negative impact. DGFSL-3D localization algorithm has high localization accuracy.