A Novel Method to Improve the Accuracy of the RSSI Techniques Based on RSSI-D

In this paper, a novel method is proposed to estimate the distance between sensor nodes based on a statistical model of GMM, which is established with the offline RSSI values, called RSSI-D. In order to estimate the arguments of RSSI-D with unobserved latent variables, EM is used to calculate the arguments of RSSI-D as accurately as possible in this paper. After RSSI-D established, the probability of sub models of RSSI-D that the online RSSI values obey can be calculated by the posterior probability in Bayesian statistics. Based on the probability of sub models corresponding to the distance segment, the distance between sensor nodes can be estimated by the weighted value of distance segments. In the simulation, the accuracy is improved obviously by RSSI-D and the arguments such as K and N are provided by the simulation results