Integration of Spectrum Database and Sensing Results for Hybrid Spectrum Access Systems

Abstract

The database-driven spectrum access system recently attracted increasing amounts of attention. It has more benefits compared to the traditional sensing-based systems. To build a more practical and reliable system, a hybrid sensing-based and database-driven spectrum access system is a promising solution. In this paper, we consider the integration problem of the database information and sensing results, which is a very important factor in order in realizing the hybrid system. We propose the integration framework, which is implemented on the database engine. The framework is divided into two main components. The first one is to process the sensing results, which contains the predictions for locations without sensing results, and the fusion policy on the sensing samples. The second component is the dynamic integration process of the generated sensing results and the database information. We first model the evaluation of the integration results as a Partially Observable Markov Decision Process (POMDP), which enables the database engine to know its current status. Then, we propose an iterative algorithm for the database engine to dynamically adjust its integration policy. In this way, the balanced status of the generated spectrum map is maintained. Simulations are conducted to reveal the performance of our framework.