Modeling and Evaluation of Remote Temperature and Conductivity Sensor Using Radio Frequency Identification Tags

Abstract: The goal of this master thesis was to model and evaluate a new RFID sensor technology developed at GE research. The sensor uses an ordinary RFID tag to simultaneously sense the temperature and conductivity of a liquid placed on the other side of a plastic membrane. This is accomplished by using a reader antenna and a network analyzer to perform a frequency sweep to calculate the impedance at the different frequencies. Because the resonant circuit properties change with temperature and the conductivity of the liquid the resulting spectrum can be used to calculate the temperature and conductivity. This is done using a mathematical method called Principal Component Analysis (PCA). The model used to describe the system is a simple circuit involving two smaller circuits connected inductively through two coils. This model was then used together with measurements to find the temperature and conductivity dependence of the different components in the model. The resulting simulation captured the bulk of the behavior of the system. PCA analysis was performed on the simulated data and compared to measurements. Considerations of noise was added and evaluated. This shows that even through there are some issues with the model it can predict certain behaviors of the sensor.