A localization method with dynamic weight argument based on Gaussian mixed Model

The Weighted Center Localization (WCL), which is a classical localization scheme for the Wireless Sensor Network (WSN), can´t decrease the average localization errors and the maximum errors simultaneously due to the fixed weight argument. Improving the localization accuracy of WCL is to decrease the distance measurement errors or to set a proper dynamic weight argument. We have already improved the distance measurement accuracy based on a statistical Gaussian mixed Model (GMM) based on the off-line Received Signal Strength Indication (RSSI) collected under the mine. To improve the localization accuracy further, we provide a dynamic weight argument based on GMM. The attributes of the dynamic weight argument emphasize the effect of the large RSSI and prevent the large RSSI from enlarging the maximum localization errors. The simulation results describe that the method proposed in this paper decreases the average localization errors and the maximum errors simultaneously.