2-Tx Alamouti Coded OFDM Systems Based on Prototype Implementation

Abstract

The progression of cellular wireless communications from 3G in the past to 4G in the present and then to 5G in the near future has been motivated by the huge demand for fast and reliable communications over the wireless channel but with lower complexity at both the transmitter and the receiver. Several impairments in the channel such as time selective fading and frequency selective fading (multipath) have been the main challenges to the wireless communications. With the advent of 4G broadband wireless communications, MIMO OFDM has been used to address the challenges. Fortunately, MIMO OFDM has proved to be one of the promising technology today. And the future of wireless communications has been shaping towards the realization of massive MIMO and millimeter (mm) Wave technologies. With the introduction of mm Wave technology, channel bandwidth will increase manifold and the severity of frequency selectivity will rise simultaneously. Hence massive MIMO OFDM may become an indispensable part of the future wireless communications. Several researches have been done on MIMO OFDM, but most of the papers quantify their results via simulation. In this thesis, we investigate the MIMO OFDM technology based on system implementation and real world experiments. Considering the lower complexity of the Alamouti scheme, we resort to the Alamouti coded OFDM systems. To be more precise, we compare the performance of two 2-Tx Alamouti coded OFDM systems based on prototype implementation. First we adopt the decision directed method proposed by Alamouti, Tarokh and Poon for single carrier systems in their paper by modifying it for OFDM system. This method employs the Alamouti encoding scheme for each subcarrier during two OFDM symbol periods and it transmits the Alamouti encoded pilot symbols in all subcarriers in first two OFDM symbol periods (Alamouti STBC). The channel coefficient of each time frequency bin is then estimated in a decision directed manner. Using these channel coefficients, simple maximum likelihood (ML) combining is used to decode the user data. This system has several drawbacks such as error propagation, inaccuracies in channel measurement and less robustness to time varying nature of the channel. So we resort to interpolation based method to realize the Alamouti coded OFDM system. This method employs the Alamouti scheme for each pair of two consecutive subcarriers within an OFDM symbol period (Alamouti SFBC). In contrast to the decision directed method, pilot symbols are not Alamouti encoded. Instead the pilot symbols are transmitted periodically in time and frequency, and the channel estimation is performed by linearly interpolating the magnitudes and phases of neighboring pilot samples. After the channel is estimated, again a simple ML combining and decoding scheme is used to decode the user data. The performance of these two systems is compared based on outdoor field test. Finally we show that the latter system outperforms the former in the same channel in terms of spectral efficiency, received SINR, error propagation and robustness to time varying channel.