Privacy Preservation Needed for Smart Meter System: A Methodology to Recognize Electric Vehicle (EV) Models

This paper introduces a practical method to determine the EV model (Car Make A model B) from high resolution (/1min) energy consumption data. The proposed method shows the importance of privacy preservation for smart meter system. The paper demonstrate the decision making process as solving a multiclass classification problem. In particular, we focus on extracting the key features of given EV charging profiles, and using the features as attributes to set up a Decision Tree (DT). We illustrate the classification problem in a 2-dimensional space and train the decision boundaries of the DT by labeled "dataid-EV model" data sets. We show that using the trained DT is efficient to predict the model of several type-unknown EVs in a distribution grid. The results would help in developing privacy-enhanced loads metering methods.