Font Size: a A A

Positioning In Wireless Sensor Network Based On Radio Frequency

Posted on:2011-09-18Degree:MasterType:Thesis
Country:ChinaCandidate:X B XuFull Text:PDF
GTID:2178330338478113Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Positioning is one of the important technologies of Wireless Sensor Network. The positioning algorithm based on Radio Frequency (RF) signal is widely used because it can realize positioning only using node RF signals itself and does not need additional hardware support. The traditional RF positioning algorithms existing problems such as positioning process complex and poor positioning accuracy, studying new RF positioning algorithm has important practical significance. This dissertation focuses on introducing the Support Vector Machine to Wireless Sensor Network RF positioning. The main contents of our study include:(1) Support Vector Regression RF positioning algorithm. For the problems of poor positioning accuracy and RF modeling difficult in traditional RF positioning algorithms, we put forward a SVR RF positioning algorithm. This algorithm obtains node positions directly by node RF signals by using SVR tools to build the positioning model whose inputs are node RF signals (RSSI and LQI for characteristics) while outputs are coordinates. This algorithm is easy to realize and can reach a high position accuracy with average positioning error about 1-m.(2) Support Vector Classification RF positioning algorithm. By classifying the positioning area into different categories, we convert positioning into a multi-classification problem based on node RSSI values to realize two indoor algorithms—the symbolic and physical positioning. We use multi-classification results as node symbolic positions, and grid center coordinates of multi-classification results as physical positions. Simulation results show that symbolic positioning is good and physical positioning can reach high position accuracy when the grid is intensive.(3) Location Fingerprint RF positioning algorithm. We improve the positioning of RADAR in two aspects: on the way, we use RSSI and LQI instead RSSI as Location Fingerprint. On the other way, we build the model between Location Fingerprint and node position rather than marring them. The improved positioning technique can not only improve position accuracy but also can realize single access point positioning.
Keywords/Search Tags:WSN, SVM, positioning, RSSI, LQI, Location Fingerprint
PDF Full Text Request
Related items