ABSTRACT
CONTENTS 15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372 15.2 System Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 15.3 Performance Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 15.4 Performance Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
15.4.1 Selected Metrics for Perfect or Unreliable Spectrum Sensing . . . . . . . . . . . . . . 385 15.4.2 Selected Metrics for Tolerable Service Degradation . . . . . . . . . . . . . . . . . . . . . . . 386
15.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Abstract
Current radio spectrum usage has been shown to be highly underutilized. Opportunistic spectrum sharing (OSS) is a promising solution for efficient spectrum utilization. In this chapter, we analytically model an opportunistic spectrum sharing system under the conditions of perfect spectrum sensing, unreliable sensing, and tolerable service degradation through queueing theoretic frameworks. The considered OSS system consists of primary users and secondary users that share a set of channels over a coverage area. Both initiating secondary users and ongoing secondary users sense the channels and perform appropriate activities. Either a buffer or infinite queues are incorporated into the proposed models for performance analysis. Sensing errors from either initiating or ongoing secondary users are considered in appropriate models, which may cause false alarm and misdetection events and impose various impacts on both types of users. We solve the steady-state probabilities of the considered systems under perfect sensing, unreliable sensing, and tolerable service degradation. We also derive a set of performance metrics of interest. Different problem Solving
methods, e.g., matrix analytic method and generating function technique, are used for solving the system equations. The proposed modeling methods are expected to be used for design and evaluation of future opportunistic spectrum sharing networks.
