| Wireless sensor networks is a kind of application-centric network, node localization problem is not only the premise of target identification, monitoring, tracking and many other applications in wireless sensor network, but also the hot issues of sensor networks. The positioning of the traditional methods requires additional hardware, or positioning accuracy is not high. Therefore, it is of great important in the wireless sensor network environment to achieve an effective, low-cost, practical use of node location algorithm to ensure performance and practicality.The positioning method based RSSI does not require additional hardware, but the RSSI susceptible to external factors (such as the environment, multi-path effects, etc). Based on the study of nodes of this interference in the direction of antenna and environment, this paper presents two kinds of interference:the variance-level interference and the mean-level interference. For these two kinds of interference, this paper presents a positioning algorithm Amendment to the RSSI from two levels:First, in the acquisition stage of RSSI, this paper makes a Gaussian model based on deviation (D-Gaussian) to improve the availability of RSSI signal; followed in the Ranging stage, this paper introduces a dynamic window mechanism EWMA and named R-EWMA to deal with the mean-level interference. At the same time, according to different applications, it can be adjusted between stability and sensitivity to broaden the algorithm applications.The experimental results show that research in this article effectively solved the conflict between the positioning accuracy and hardware costs in the traditional algorithm, the processing to RSSI can fundamentally reduce the introduction of error. At the same time, R-EWMA algorithm can further reduce the impact of RSSI on the positioning accuracy from ranging level. Of course, at the time to improve positioning accuracy, the algorithm has introduced the larger delay and computational complexity. |
PDF Full Text Request