Beamforming And Large-Scale Multi-antenna Transmission Technologies For Next-Generation Wireless Communications Of High Speed Railway

The railway transportation especially the high-speed railway (HSR) transportation has attracted a lot of attention in the world due to its obvious advantages such as high speed, large transport capacity, little environmental pollution, low energy consumption, good safety and comfort. HSR is a comprehensive concept which refers to the railroad track and the power supply network and contains the high-speed train as well as the communication and signal systems. It usually operates at the speed of 200km/h to 350km/h, the experimental peak value of which is as high as 574km/h. The support of various kinds of modern radio and wire communications and information processing techniques is vital for the secure and efficient operation of HSR. The wireless communication of HSR is mainly composed of two parts:the dedicated wireless communications of train dispatch and control and the broadband wireless access for passengers. The current wireless communication systems of HSR employ GSM-R to satisfy the system requirement of train dispatch and control, which is based on Global System for Mobile communication (GSM) and designed specificly for the railway. However there is temporarily no specified wireless communication system providing information service for passengers, and they still use the UMTS networks based on the 3rd Generation of mobile communication (3G) or LTE based on so-called 4th Genaration of mobile communication (4G) operated by the public wireless carriers. To fulfill the demand of next-generation of wireless communication systems of HSR for high reliability and large capacity, GSM-R is evolving to Long Term Evolution for Railway (LTE-R) while the broadband wireless access for passengers also needs a more effective solution. Thus, in this thesis, we will focus on the high-performance wireless transmission technologies under HSR scenario, including the research and design of beamforming techniques and large-scale multiple antennas (LSMA) system.Firstly, since in HSR scenario the location and speed information can be shared by the track-side evolved NodeB (eNodeB), the traditional beamforming technique which is mainly used to reduce the interference and increase the signal-to-noise ratio (SNR) of desired user during the transmission, is capable of assisting the handover process of the transceiver deployed on the train. The reliability of handover will be improved by adjusting the beamforming gains. The results show that the proposed scheme enables the train to be served by the beams from source eNodeB and target eNodeB simultaneously which is beneficial to the handover trigger. It eliminates the risk of outage and improves the quality of received signal in order to realize the reliable handover.Secondly, through extending the single onboard transceiver to multiple sigle-antenna transceivers, the opportunistic beamforming (OBF) technique can be employed, as OBF based on the dumb antennas only requests the feedback of channel SNR while the smart-antenna beamforming has higher requirement on channel estimation. OBF can increase the dynamic range of channel fluctuations between the antennas of eNodeB and onboard transceivers by multiplying random complex weight to each transceiver antenna. Furthermore, the random phase of the complex weight can be affected with the predictable train location information, which will help obtaining more multi-user diversity gain. The results show that the proposed scheme can improve the system throughput and make the performance of the conventional OBF approach the performance of the coherent beamforming technique based on smart antennas.Thirdly, it is necessary to redesign the structure of transmit and receive antenna arrays for combining the beamforming technique and spatial diversity technique in order to get array gain and diversity gain, as the beamforming technique requires the antenna spacing to be smaller than half the carrier wavelength while the antenna spacing has to be five to ten times of the carrier wavelength when the Alamouti space-time block coding (STBC) is applied. The combined schemes are introduced to HSR scenario, including the transmit beamforming (TxBF) and receive beamforming (RxBF). Then they can be categorized as STBC-TxBF Ⅰ and Ⅱ structure as well as STBC-RxBF Ⅰ and Ⅱ structure, in which the STBC-RxBF Ⅱ is the proposed structure. STBC-RxBF Ⅱ generates the weights based on the direction of arrival (DOA) instead of the eigenvalue decomposition (EVD) of the channel covariance matrix and forms two beams to receive the spatial streams processed by STBC at the transmitter. The results show that the proposed scheme can decrease the algorithm overhead and attain the diversity and array gain simultaneously to improve the performance of bit error rate (BER) of the combined scheme.Then, although the beamforming can be combined with the spatial diversity technique, the problem of high channel correlation caused by strong line-of-sight (LOS) path in HSR scenario still needs to be solved when it is combined with the spatial multiplexing technique. In this thesis, the high angular-resolution dual beams which are the feature of large-scale multiple antennas (LSMA) technique are used to serve the transceivers deployed on the head and tail of the train respectively to get array gain and multiplexing gain. As the inter-beam ambiguity increases when the train moves towards the cell edge, and the angular resolution of the beam is decided by the number of active antennas as the eNodeB, a transmission strategy based on adaptive antenna activation (AAA) is proposed to switch between different beamforming schemes according to the location of the train. The results show that the proposed AAA based beamforming scheme can obtain the multiplexing gain in HSR scenario and maintain good and stable performance of spectral efficiency and BER while the train moves through the entire cell.Lastly, the LSMA can be deployed either centrally or distributedly. In HSR scenario, the high angular-resolution beams generated by activating all the antennas at the eNodeB can assist the transmission significantly with centrally located LSMA, but the path loss will be detrimental to activate all the antennas if the distributedly located LSMA is applied. In this case, the antenna selection is needed. After deriving the performance upper bound for the antenna selection algorithm of distributed LSMA, a weighting pre-processing method for time-varying channels is proposed to improve the spectral efficiency of antenna selection algorithms in HSR distributed LSMA communication systems.