Electric vehicle fingerprinting: identification through charging behavior

Abstract

As plug-in electric vehicles (PEVs) become increasingly common, they are expected to become highly integrated with electric power generation and distribution systems. If these systems are to support PEVs’ demands for power, PEV charging will require substantial coordination. A large body of research has developed proposing strategies and technologies to achieve this coordination. While some of these approaches have been implemented, there is no standard system for PEVs and parties coordinating charging to exchange key information, such as vehicle parameters. This thesis describes an experiment to enable electric vehicle supply equipment to obtain key parameters by fingerprinting the PEV over the charge cord. The results of the experiment were promising, with the best classifier (a random forest) constructing a model that correctly classified 100% of the PEVs in the somewhat limited dataset.