Research Of Fault-Tolerance Target Localization Based On Wireless Sensor Networks

In wireless sensor networks, it is meaningless for the data without location information. Target localization is an important application for the sensor networks. The difficulty it faces is the uncertainty of the target position. Since the data collected by the nodes is inaccurate which is caused by the instability of the sensor network itself, how to overcome the faulty and unstable factor so as to improve the accuracy of the target localization has an important significance. This thesis studies the fault-tolerance target localization in wireless sensor networks, improving the fault tolerance of the algorithm and the localization accuracy.Firstly, the research status at home and abroad are deeply analyzed and summarized. Technical progress and key issues are discussed about the target localization in wireless sensor network. Furthermore, aiming at robust localization in wireless sensor networks, this thesis proposes a new modified localization algorithm. On the basis of the relationship between the signal attenuation model and the distance, we divide the received signal strength into ten different levels from zero to nine. The node determines which level it is in according to the received signal strength from the target, divides its coverage area into grids and gives a grid weight to each grid. By accumulating the data of each grid, we can find that the position of the target locates at the grid which corresponds to the maximum value. The simulation results show that the proposed algorithm, which changes the coverage area of the estimated target and the grid value of the grids dynamically according to the received signal strength, is simple and efficient to carry out. It can effectively improve the localization accuracy and has a better fault tolerance.Taking into account that the multi-target localization algorithm is easy to fall into local optimum situation, this thesis uses K-means algorithm to divide the sensor nodes into a certain group which is roughly corresponding with the target. Then a multi-target fault detection model is introduced based on false alarm rate, behind which the tabu search algorithm is used to locate the targets. Simulation results show that the algorithm in this thesis has good fault tolerance. The K-means algorithm can separate the nodes with a high accuracy which effectively avoids falling into local optimum. Compared with the expectation maximization algorithm (EM algorithm) and multi-resolution algorithm (MR algorithm), the algorithm in this thesis reaches a higher positioning accuracy.Considering the effects of non-line of sight (NLOS) error on the positioning accuracy, this thesis proposes a residual weighted algorithm to reduce the NLOS error. By gradually reducing the number of the distance measurement combination, a set with the smallest residual is selected for the final location by weighted average method. Simulation results show that the algorithm can effectively reduce the influence of the NLOS error. Meanwhile, the algorithm still has a better accuracy with the increase of the NLOS error and NLOS nodes. At last, the results prove that the algorithm proposed in this thesis has a good robustness under NLOS environment.At last, the research in this thesis is summarized. Future research work is proposed.