Cross-layer design and optimization of heterogeneous cellular mobile networks

The rapid growth of wireless communication and access, in conjunction with increasing demand and sophistication of wireless applications, supplicate intelligent and reliable systems to support the exchange of large classes of traffic with rising quality of service (QoS) requirements. In addition, co-operation among cellular systems that incorporate different radio access technologies is notably important as the current third generation (3G) universal mobile telecommunication systems (UMTS) are expected to co-exist with the emerging fourth generation (4G) long term evolution (LTE) technologies for years to come. To these ends, radio resource management (RRM) techniques across different network layers, conjointly with spectrum sharing strategies, are vital in achieving desirable performance in heterogeneous networks. In this thesis, novel cross-layer design strategies, for jointly optimizing the physical (PHY)-layer and data link layer (DLL) parameters, are proposed in the contexts of code division multiple access (CDMA) and shared-spectrum heterogeneous orthogonal frequency division multiplexing (OFDM)/CDMA networks. These strategies facilitate dynamic radio resource allocation by exploiting the random variations of channel and network activity. Transmit power and QoS constraints are imposed on systems to maintain communication costs, effectiveness and quality. This thesis makes four main contributions. Firstly, in the proposed cross-layer techniques, based on automatic repeat request (ARQ) delay limits and prescribed maximum packet loss rates in the DLL, the optimum outer-loop power control (OLPC) SNR-targets and the corresponding adaptive spreading factors are derived in the PHY-layer, as functions of the number of active users in the cell. The optimality is, in this sense, maximization of cell effective throughput. Secondly, the performance of the proposed interference-based resource allocation schemes are evaluated over Nakagami-m frequency -flat and -selective fading channels. In particular, frequency-selective channels with maximum ratio combining (MRC) RAKE receiver are considered. Thirdly, I consider the uplink in multi-user cellular communication systems with single- and multi- service traffic scenarios, which are respectively modelled with one- and multi- dimensional discrete Markov chains. Finally, a dynamic cross-layer resource allocation algorithm in the context of shared-spectrum heterogeneous OFDM/CDMA networks is proposed. Opportunistic spectrum access (OSA) is employed to utilize the idle parts of the primary spectrum, effectively minimizing the interference levels, to maximize the total deliverable secondary throughput. Throughput performance of the optimized schemes and the achieved improvements, relative to the non-optimized and state-of-the-art schemes, are demonstrated with theoretical and simulation results for various settings of system parameters.