Spectrum sharing in wireless cellular networks

• Main Author: Han, Shi Ying
• Other Authors: Soong Boon Hee
• Format: Theses and Dissertations
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Wireless communication systems
• Online Access: https://hdl.handle.net/10356/63699
• Institution: Nanyang Technological University

To support the explosive growth of wireless services, wireless cellular networks are facing escalating spectrum scarcity problem nowadays. Fixed and exclusive allocation of radio spectrum among telecom operators and generations of cellular technologies, which is widely adopted currently, has been proven to be inefficient in utilizing the radio resource. In particular, as cellular technology evolves, the frequency bands allocated to the new generation cellular networks are getting crowded, while those maintained by legacy cellular networks become under-utilized. Thus, by allowing the new generation cellular networks to dynamically access to the licensed bands of legacy cellular networks has attractive capability in improving spectrum utilization and mitigating spectrum scarcity. Such demand of spectrum sharing gives birth to spectrum refarming (SR) technique which is a new paradigm of spectrum sharing of cellular networks. Since Orthogonal Frequency Division Multiple Access (OFDMA) is the major radio access technique for the new generation cellular networks, the spectrum sharing between OFDMA and Code Division Multiple Access (CDMA) or Global System for Mobile Communications (GSM) systems is of most interest here. In this thesis, we mainly focus on the SR technique for OFDMA/CDMA coexistent system with different types of infrastructure sharing, including active infrastructure sharing, passive infrastructure sharing and non-infrastructure sharing. First, a fundamental SR model is constructed with active infrastructure sharing where the OFDMA and CDMA systems operate in the common frequency band with sharing of cell site and base station (BS) antenna. The user signal-to-interference-plus-noise-radio (SINR) is quantified based on which the interference margin provided by CDMA system is predicted by OFDMA system. With the interference margin, the OFDMA system allocates power and subcarrier resource efficiently with protection to the CDMA system. Relying on the CDMA inner power control, the spectral efficiency of the SR system is further improved by exploiting the interaction between OFDMA and CDMA systems. Next, the fundamental SR model is extended into passive infrastructure sharing model where the OFDMA and CDMA systems share the common cell site but adopt separate antennas at BS. This infrastructure sharing would be popular, as it can achieve higher SR throughput compared with the active infrastructure sharing. A realistic and crucial problem is that the channel power gain from secondary transmitter (STx) to primary receiver (PRx) is hard to know due to the lack of direct communication between the two systems in practice. To solve this problem, we propose a novel resource allocation method for OFDMA system. Somewhat surprisingly, this method can sufficiently protect the CDMA system despite the channel power gain aforementioned is unavailable. Furthermore, utilizing the CDMA inner power control, two innovative resource allocation schemes are proposed for OFDMA system that farthest exploit the sharing opportunity provided by CDMA system. Without direct signalling between the two systems, OFDMA system can compel CDMA system to operate at the optimal status by controlling its own transmission strategy. Finally, we relax the SR model from infrastructure sharing where the two systems share no infrastructure. This naturally invokes the SR application in the heterogeneous networks where multiple OFDMA small cells share the licensed spectrum of CDMA macrocell network. We quantify the interference between the two systems, based on which the OFDMA resource allocation problem is formulated and solved efficiently. The problem of unavailable channel power gain from STx to PRx is also considered for which solution is proposed.