Researches And Applications In Localization Of Wireless Sensor Network Based On RSSI

Wireless Sensor Network (WSN) is growing with the development of MEMS(micro-electro-mechanism system), wireless communication and digital technology. WSN will change interactive mode between human and the objective world. As an important application in WSN, localization technology is concerned by researchers for a long time. Based on current localization technologies, the main tasks of this paper are research of localization technologies using Received Signal Strength Indicator (RSSI) in WSN.First, this paper analyses the situation of current localization technologies in WSN, and indicates the advantages and disadvantages of them.Then, because the random error of RSSI in NMDS-RSSI ,which based on Non-Metric Multidimensional Scaling(NMDS), is great, so maximum likelihood estimation is introduced. Error is effectively reduced through calculating average value after several measures. This paper proposes NMDS-MLE-RSSI algorithm, and network communication model for it. The application background is analyzed.And then, this paper simulates NMDS-MLE-RSSI with MATLAB to verify its performances. The results indicate that NMDS-MLE-RSSI is better for application of indoor environment. Localization results have certain distortion phenomenon when NMDS-MLE-RSSI is employed in mine tunnel environment, which is long and narrow.In addition, RSSI value is easy to jump in practical. Aimed at mine tunnel application environment, This paper weights the coordinate of anchor(position known) node with RSSI between anchor and position-unknown node, then calculates the coordinate of position-unknown node by average weighted coordinates of several anchors. This paper proposes a RSSI-Weighted Centroid localization algorithm. This paper do experiments in a mine tunnel liked environment with Texas Instruments Z-Stack protocol, CC2430 nodes developed in our laboratory and the PC Software Sensor-View. The result of localization performance is good.