Hierarchical and Hybrid: Mobility-Compatible Database-Assisted Framework for Dynamic Spectrum Access

Abstract

The protection of primary user activities plays a vital part in dynamic spectrum access. There are currently two types of schemes. One is based on spectrum sensing and the other one relies on spectrum database. There exist many works surrounding the subject of spectrum sensing, which requires high sensing accuracy, and thus poses extra time cost. Nevertheless, the database-based access scheme is receiving an increasing amount of devotion, e.g., IEEE 802.11af for TV white space. In comparison to spectrum sensing, nodes only need to look up the spectrum maps provided by the database. In this paper, we study the practical issues of the above two schemes, and propose a hierarchical framework, which enables the hybrid spectrum access scheme. In this framework, we build relatively reliable clusters, and have cluster heads connect to the spectrum database to assist nodes in their clusters. As a result, nodes with poor or no connections to the database can benefit from spectrum maps as well. The process of retrieving spectrum maps is formalized as a Markov Decision Process. Moreover, mobility compatibility is provided. We illustrate how the evolution of our hierarchical structure is affected by mobile nodes. We also propose a feasible algorithm to maximize the benefits that can be obtained from the database under the mobile environment. We model a virtual database and conduct simulations to reveal certain performances of our framework.