Prediction of RFID systems coverage applied to smart cards scenario

RFID systems usually need to be applied where the installation cost can be very high, and thus, running a simulation in such environments would bring many benefits such as saving time and money. In this case, the simulation could be considered as a prediction method that verifies whether the configuration for the desired scenario works correctly and meets the system's requirements. This work presents a research about statistical prediction models for RFID systems' coverage in a specific scenario. In order to predict the RFID system coverage, the tag read count was measured in an indoor environment. In the chosen scenario, RFID technology is applied to smart cards. This research presents a model using Multiple Linear Regression and another approach using Artificial Neural Networks on the prediction of the read count. Results showed improvements on the prediction of area coverage, bringing advantages to RFID projects, such as less time and resources required to RFID systems deployment.