Cyberthreat Analysis and Detection for Energy Theft in Social Networking of Smart Homes

The advanced metering infrastructure (AMI) has become indispensable in a smart grid to support the real time and reliable information exchange. Such an infrastructure facilitates the deployment of smart meters and enables the automatic measurement of electricity energy usage. Inside a community of networked smart homes, the total electricity bill is computed based on the community-wide energy consumption. Thus, the coordinated energy scheduling among smart homes is important since the energy consumptions from some customers can potentially impact bills of others. Given a community of networked smart homes, this paper analyzes the energy theft cyberattack, which manipulates the energy usage metering for bill reduction and develops a detection technique based on Bollinger bands and partially observable Markov decision process (POMDP). Due to the high complexity of the POMDP-solving process, a probabilistic belief-state-reduction-based adaptive dynamic programming technique is also designed to improve the detection efficiency. Our simulation results demonstrate that the proposed technique can successfully detect 92.55% energy thefts on an average while effectively mitigating the impact to the community. In addition, our probabilistic belief-state-reduction-based adaptive dynamic programming technique can reduce the runtime by up to 55.86% compared to that without state reduction.