On Multiantenna Spectrum Sensing and Energy-Aware Resource Allocation in Cognitive Radio Networks

Abstract

This dissertation covers two main parts: 1) low complexity cyclostationarity based spectrum sensing in cognitive radio, and 2) energy efficient resource allocation in cognitive radio networks. Cognitive radio allows unlicensed users (secondary users) to access the frequency spectrum allocated to licensed users (primary users). Secondary users are allowed to access the spectrum licensed to primary users in two ways: underlay and overlay. Under the overlay mode, secondary users need to sense the channel first and then decide whether to transmit based on the presence or absence of a primary signal. The underlay mode allows primary and secondary users access to the same channel simultaneously while constraining the transmitted power of secondary users so that it can be treated as background noise at primary users without exceeding the primary users noise floor. The most common sensing method is based on the energy distribution of the primary signal, however its performance is degraded when noise variance is not accurately known. For cyclostationary primary signals such as Orthogonal Frequency Division Multiplexing (OFDM) and Gaussian filtered Minimum Shift Keying (GMSK) signals, we exploit their statistical characteristics under both certain and uncertain Gaussian noise and propose a low complexity cyclostationarity-based spectrum sensing method for cognitive radio networks. Multiple receiving antennas are utilized to improve sensing performance. In addition to sensing complexity, energy efficiency has become a “hot” research topic due to the trend of sustainable energy and green wireless communication and networks. For mobile terminals with battery for power supply, the energy consumption issues are more serious because development of battery technology lags behind the development of wireless communication technologies. This dissertation also addresses energy efficient sensing and power allocation for multichannel overlay cognitive radio networks (CRN), energy efficient precoding for MIMO (multiple input multiple output) assisted spectrum sharing CRN and MIMO cognitive multiple access (C-MAC) channel.

The dissertation is organized into eight chapters. Chapter 1 provides an introduction to cognitive radio networks and related research issues. Existing spectrum sensing algorithms are addressed too. It also briefly outlines the whole dissertation. Chapter 2 through Chapter 4 are mainly concerned with low complexity cyclostationarity-based spectrum sensing for cognitive radio networks. This sensing scheme is explicitly designed for cyclostationary primary signals under white and colored Gaussian additive noise. Chapter 2 introduces the cyclostationarity background and related spectrum sensing algorithm. The cyclostationary characteristics of the OFDM and GMSK signals are illustrated as examples. The hypothesis testing framework is briefly described and existing spectrum sensing tests based on cyclostationarity are presented. In Chapter 3, a novel low complexity cyclostationarity-based spectrum sensing approach under white Gaussian noise is proposed. The theoretical performance and computational complexity advantages compared with other counterparts are given in this chapter. The simulation examples support the theoretical analysis. Chapter 4 focuses on cyclostationarity-based spectrum sensing under colored Gaussian noise. Besides sensing performance, energy efficiency related with sensing and transmission for either overlay or underlay cognitive radio networks is also studied in Chapters 5, 6 and 7. Chapter 5 focuses on energy-aware sensing and power allocation for multi-channel transmission of a single secondary link in overlay cognitive radio networks. The optimization algorithms for sensing duration, test threshold and transmission power allocation are described. Simulation results are provided in support of the proposed algorithm. Chapters 6 and 7 address energy efficient resource allocation in underlay cognitive radio networks under both transmit power constraint and interference power constraint. Chapter 6 proposes an energy efficient transmit covariance matrix (precode) to maximize energy efficiency for a single secondary link in spectrum sharing underlay CRN. Multiple antennas are employed at the secondary transmitter and receiver. Chapter 7 further extends the work in Chapter 6 to propose an energy efficient precoding for MIMO cognitive multiple access channels. Conclusions and future work are stated in Chapter 8.