Management Scheme and Optimal MAP Detection Threshold of Packet-based Cognitive Radio System

Abstract

A cognitive radio system (CRS) senses the radio spectrum environment and adaptively use unoccupied spectrum resource. The CRS should optimize between two conflicting goals for different applications. One is the minimization of interference to primary users and the other is the maximization of usage of spectrum opportunities of the secondary users. In this paper, we propose a dual-level management scheme that performs spectrum analysis and access decision making at both network level and terminal level. Using this scheme, secondary users obtain the distribution of the silent period and control their sensing performance. A closed-form optimal threshold is obtained using the maximum a posteriori (MAP) rule based on the distribution of the silent period of primary users. Furthermore, applying this optimal detection threshold, we set up an analytical model to study the tradeoff between protection to primary users and spectrum usage of secondary users. Results show that the optimal threshold changes according to the duration of sensing. Furthermore, increasing the duration of sensing might not necessarily improve the protection to primary users. Using this analytical model, the scheme chooses sensing period conditioned on the distribution of silent period of licensed users to improve the usage of unoccupied spectrum while limiting the collision between primary users and secondary users for different application purposes. I. INTRODUCTION