Mobile user localization in wireless sensor network using grey prediction method

Knowing the position of mobile user is an important role for location services in the building. The characteristics of wireless sensor network are low power, low cost and low complexity. With these functions, wireless sensor network have great potential to develop indoor position system. However, radio signal propagation is easily affected by diffraction, reflections, and scattering of radio in the building, the received signal strength need good calibration method to improve the accuracy of position estimation system. In this paper we use grey prediction method in wireless sensor network and employ wireless LAN medium (Zigbee/802.15.4). The grey prediction is used to predict the tendency of RSSI (received signal strength indicator), and we also designed dynamic triangular (DTN) location method. We have done some experiments and compare with other classical location finding methods. The mean distant error of RSSI on mobile user can be within 2.3 m at offline stage. As a result, grey predication with DTN provides more accurate predicted position and carries out mean distance error within 1.3 m at run-time stage.