Research on Localization Technology Based on UKF for WSNs
|School||Central China Normal University|
|Course||Applied Computer Technology|
|Keywords||Localization Improved Centroid Unscented Kalman Filtering WirelessSensor Networks|
Wireless sensor networks have many features, including a large degree of freedom, more number of nodes and the network layout complexity, as well as a wide range of applications, broad prospects which make it as a modern network technology research focus. Localization of wireless sensor networks is an important and basic part to play in the framework. Therefore, proposed a localization algorithm which its error is small, has low energy consumption and is easy to deploy high-quality is particularly important and urgent.Currently existing localization algorithms widespread over-reliance on the environment, large power consumption, low accuracy and poor usability. This paper presents two algorithms can be well alleviate these problems, which are improved centroid localization algorithm based on common centroid algorithm, it has a smaller error and higher precision, and the UKF localization algorithm; The UKF algorithm uses the results of improved centroid algorithm as the initial value of the unknown node, on this basis, combine with UKF theory to achieve higher accuracy. The main research contents are as follows:(1) An improved centroid localization algorithm is proposed based on the common centroid localization algorithm which only uses geometric centroid. But the improved one uses the idea of signal path loss model that shows the closer distance between nodes, the smaller uncertainty, as the theoretical background for positioning. Simulation results show that the algorithm compared with ordinary centroid algorithm, has smaller average error and maximum error as well as higher accuracy.(2) To further improve the accuracy of the centroid localization algorithm, the UKF algorithm applies its results as the initial values and combines with filtering technology to increase the precise of nodes. Because the established algorithm model is nonlinear, it’s necessary to adopt nonlinear filtering techniques to handle it, the UKF is apposite to deal with nonlinear estimation problem using unscented transform which can transfer the statistical properties of nonlinear systems. The simulation results show that the localization algorithm based on UKF has improved the accuracy compared with centroid and EKF localization algorithm, simultaneously, it has good stability.The two proposed localization algorithms in this paper both have better location accuracy and good stability. Before localization, we can preset a specific threshold value of accuracy, if improved centroid localization algorithm’s results meet the accuracy requirements, to end directly, or we need continue to use UKF localization algorithm for precise.