Privacy-Cost Management in Smart Meters With Mutual-Information-Based Reinforcement Learning

The rapid development and expansion of the Internet Things (IoT) paradigm has drastically increased collection exchange data between sensors systems, a phenomenon that raises serious privacy concerns in some domains. In particular, smart meters (SMs) share fine-grained electricity consumption households with utility providers can potentially violate users’ as sensitive information is leaked through data. order to enhance privacy, consumers exploit availability physical resources such rechargeable battery (RB) shape their power demand dictated by privacy-cost management unit (PCMU). this article, we present novel method learn PCMU policy using deep reinforcement learning (DRL). We adopt mutual (MI) user’s load masked seen grid reliable general measure. Unlike previous studies, model whole temporal correlation MI its form use neural network estimate MI-based reward signal guide process. This approach combined model-free DRL algorithm known double ${Q}$ -learning (DDQL) method. performance complete DDQL-MI assessed empirically an actual SMs set compared simpler measures. Our results show significant improvements over state-of-the-art privacy-aware shaping methods.