Location Estimation for Wireless Sensor Retrieval

Date Added: Apr 2010
Format: PDF

Wireless sensor retrieval is importance for cost saving, data analysis and security purposes in WSNs. Finding sensors is especially challenging because both the number and locations of these sensors could be unknown. In this paper, the authors propose an online probabilistic localization algorithms that iteratively calculate the locations of multiple wireless sensors in WSNs, called LEGMM (Location Estimation based on Gaussian Mixture Model). In each iteration, they figure out part of the sensor fields, in which they first identify coarse-grain sensor locations by maximizing the likelihood estimate over a grid structure, called Grid-LEGMM (Grid Location Estimation based on Gaussian Mixture Model), then use the EM-method to refine the results of Grid-LEGMM, called EM-LEGMM (Expectation Maximization based on Gaussian Mixture Model).