System-Level Modeling and Optimization of MIMO HSDPA Networks

Interaction between the Medium Access Control (MAC)-layer and the physical-layer routines is one of the basic concepts of modern wireless networks. Physical-layer dependent resource allocation and scheduling guarantee efficient network utilization. Accordingly, classical link-level analyses, focusing only on the physical-layer are not sufficient anymore for optimum transceiver structure and algorithm development. This thesis presents the development and application of a system-level description suitable for the downlink of Multiple-Input Multiple-Output (MIMO) enhanced High-Speed Downlink Packet Access (HSDPA with particular focus on the Double Transmit Antenna Array (D-TxAA) transmission mode. The system-level model allows for investigating and evaluating transmission systems and algorithms in the context of cellular networks. Two separate models are proposed to obtain a complete system-level description: (i) a link-quality model, analytically describing the MIMO HSDPA link quality in a so-called equivalent fading parameter structure, and (ii) a link-performance model, capable of predicting the BLock Error Ratio (BLER) performance of the transmission. Based on this novel system-level model, a flexible system-level simulator is developed. The simulator, implemented in Matlab, is subsequently applied to a set of different optimizations. First, some general investigations of the system behavior of D-TxAA MIMO HSDPA in the cellular network context are conducted. This leads to recommendations and guidelines for network planning. Afterwards a Radio Link Control (RLC) based D-TxAA stream number decision algorithm with enhanced robustness against outdated User Equipment (UE) Channel Quality Indicator (CQI) feedback is developed. Similarly, a novel content-aware Medium Access Control for High-Speed Downlink Packet Access (MAC-hs) scheduler increasing the Quality of Experience (QoE) for video-streaming is proposed and its performance is assessed. Besides the research focusing on the High-Speed Downlink Shared CHannel (HS-DSCH) also the signaling channels?being very important for the operation of wireless networks?are investigated. In particular, an optimization of the Common Pilot CHannel (CPICH) power allocation for various network configurations is conducted. The findings are especially important for MIMO HSDPA trials, but can also be valuable for the optimization of already existing MIMO HSDPA cell clusters. Finally, a multi-user interference-aware Minimum Mean Squared Error (MMSE) equalizer is developed. This novel receiver takes the structure of the Transmit Antenna Array (TxAA) interference in the cell into account. The underlying system-model represents an extension to the classical single-user analysis. The theoretical limits and the throughput are evaluated by means of physical-layer and system-level investigations under realistic channel conditions. The results show significant performance gains of the interference-aware equalizer for channels with large delay spread.