Gene Expression Programming Model for Determining Location in Wireless LANs

The traditional path loss model of log distance assumes that the distance between users and access points has the exponential decline relationship with the received signal strength. However, a large number of experiments show that the model is not accurate. Therefore, the relationship between the distance and the RSS can not be summarized by a simple function. We can investigate the hidden principle of data by all kinds of evolutionary algorithms such as gene expression programming or GEP in order to further improve the accuracy of the model. In this paper, an indoor positioning model based on gene expression programming or GEP is proposed, and it makes full use of GEP to dig up the hidden non-linear relationship between the distance and received signal strength. Experimental results demonstrated that the proposed GEP model can improve the average error of 20.1 meters of traditional path loss model to 4.4 meters when using trilateration method.