Context-Aware Learning-Based Resource Allocation for Ubiquitous Power IoT

Abstract

Ubiquitous power Internet of Things (UPIoT) will revolutionize every aspect of the energy sector due to its powerful sensing capability and ubiquitous connectivity. However, the tension between the exponentially increasing number of devices and the limited spectrum resources poses a new challenge on channel selection. In this article, we propose a context-aware learning-based channel selection framework, which can learn the optimal long-term strategy without prior knowledge of global state information. Specifically, we propose a service reliability aware, energy aware, and data backlog aware (SEB) upper confidence bound (UCB)-based channel selection algorithm named SEB-UCB to address the non-adversarial channel selection problem, and propose an SEB exponential weight algorithm for exploration and exploitation (EXP3)-based channel selection algorithm named SEB-EXP3 to address the adversarial channel selection problem. Next, a case study is provided to demonstrate the feasibility of the proposed framework. Finally, we conclude this article and identify several future research issues.