Baseband Synchronization Algorithms In Multi-Antenna OFDM Wireless Communication Systems

The rapid development of information technology and the explosive growth ofmultimedia data communication create both opportunities and challenges for futurewireless communication technologies. The challenges include high demands formobility, real-time and user experience, etc. Wireless communication systems basedon multi-input multi-output (MIMO) antenna developed in recent years aredemonstrating its technology superiority. Meanwhile, orthogonal frequency divisionmultiplexing (OFDM) has been used widely in terrestrial wireless communicationsdue to its robustness against multi-path fading and its high spectral efficiency. Withthe combination of these two technologies, MIMO-OFDM based wirelesscommunication system becomes a very important area for future development ofwireless communication technology.The performance of baseband synchronization algorithm has a high impact onreliably receiving the data in a wireless communication system. The OFDMtechnology imposes a strict requirement on orthogonality between sub-carriers. Thetraditionally used cyclic prefix (CP) protection technique can partially mitigate thebaseband synchronization problem caused by sample timing error in a single antennaOFDM system. However, in MIMO-OFDM systems, multiple antennas can causeinter-antenna interference in baseband signal processing. As a result, basebandsynchronization is one of the urgent problems that need to be addressed both in theoryand in practice for MIMO-OFDM wireless communication systems.In this dissertation, timing synchronization and frequency synchronization inMIMO-OFDM wireless communication systems are studied. The theoreticalperformance of synchronization algorithms based on multi-antenna combining isanalyzed. Three improved algorithms are proposed including maximum ratiocombining timing synchronization algorithm, non-coherent sequence detectionalgorithm, and algorithm for carrier frequency offset and sampling frequency offsetestimation. These innovative algorithms are verified through link level simulationsand FPGA implementation. The main contribution of the dissertation can be describedas follows:1. The effects of synchronization errors on the performance of OFDM systemsand MIMO wireless channel models are analyzed first, which is laid as the theoretical foundation for the design of synchronization algorithms in the remaining chapters.2. A maximum ratio combining timing synchronization algorithm usingimproved signal-to-noise ratio estimation is proposed after timing synchronization isstudied for multi-antenna signals. The theoretic analysis and the performance of thetiming synchronization algorithms based on linear diversity combining are provided.The simulation results indicate that the timing synchronization performance of theproposed algorithm can be improved by about0.1dB under four antennas scenario,and is about0.5dB better than that of equal gain combining timing synchronizationalgorithm.3. An improved non-coherent sequence detection algorithm througheliminating sample timing error and smoothing the frequency domain characteristicsof the sequence is presented. This algorithm achieves the same performance ascurrently used coherent sequence detection algorithm based on imperfect channelestimation at high signal-to-noise ratio, and has a performance improvement of about0.8dB at low signal-to-noise ratio.4. A maximum ratio combining carrier frequency offset and samplingfrequency offset joint estimation algorithm using improved signal-to-noise ratioestimation is proposed after the frequency synchronization is studied at the acquisitionstage of the multi-antenna signals. Meantime, the theoretic performance of thefrequency synchronization algorithms based on linear diversity combining is provided.It is demonstrated through simulation that the improved algorithm can achieve about0.1dB performance gain under four antennas scenario and is closer to the theoreticalperformance bound. Since the performance of the proposed frequencysynchronization algorithm is about0.8dB better than that of the equal gain combiningone, it can achieve higher estimation accuracy at acquisition stage.5. To solve the problem of high area and power usage when implementing themulti-antenna synchronization algorithm using integrated circuits, a method to reducethe memory size is presented with significant improvement in area (79%reduction)and power (75%reduction) of the digital integrated circuits. The proposedimplementation method can also reduce the power of front-end analog circuitsthrough selecting different analog-to-digital converter (ADC) with different powerusage at different stage of synchronization state machine.In this dissertation, three improved innovative algorithms are proposed forMIMO-OFDM wireless communication systems based on theoretical research and performance analysis of baseband synchronization algorithm. Meanwhile, a referencedesign is provided for the implementation of the proposed algorithms using integratedcircuits.