Pilot Pattern Optimization for Doubly-Selective MIMO OFDM Transmissions
Current wireless transmission systems are far from their theoretically achievable performance bounds. The main reason behind this is a conservative approach of the standardization organizations. Most current standards for wireless communication employ Multiple-Input Multiple-Output (MIMO) Orthogonal Frequency-Division Multiplexing (OFDM) modulation as it offers a high spectral effciency. These systems require the insertion of at the receiver known symbols in order to estimate the transmission channel. These so-called pilot-symbols consume available resources such as power and bandwidth, and therefore eff ectively decrease spectral effciency. This thesis deals with pilot pattern optimization for MIMO OFDM transmission systems. First, an optimal power distribution among pilot- and data-symbols is considered. The post-equalization Signal to Interference and Noise Ratio (SINR) is maximized in order to deliver optimal performance. The optimal power o set between the pilot- and data-symbols depends on the ratio between the number of pilot- and data-symbols, and on the distinct performance of the utilized channel estimator. The achievable gains by the optimal power distribution depend on the operational point. Throughput gains up to 10% can be achieved. Furthermore, this thesis proposes a framework for optimal pilot pattern design for MIMO OFDM systems under doubly selective channels. An upper bound of the constrained channel capacity including channel estimation errors is provided. This allows to find an optimal pilot pattern for a given Signal to Noise Ratio (SNR channel correlation, and channel estimator. Significant throughput gains can be achieved by employing the optimal pilot patterns compared to transmission systems with standardized fi xed pilot patterns. The throughput gains can reach up to 850% when comparing with a 4 4 Long Term Evolution (LTE) system. In this thesis, I propose solutions how to approach the theoretically achievable performance bounds. The proposed solutions can easily be implemented into the future standards for wireless communication, and significantly improve their throughput.