BER Analysis And Sounding Optimization For TDD LTE/LTE-A Systems With Imperfect Channel State Information At Transmitter

MIMO is one of the key physical layer technologies to implement performance targets ofLTE and LTE-Advanced. There are three basic MIMO technologies: spatial diversity, spatialmultiplexing and beamforming. This dissertation focuses on the precoding-based spatialmultiplexing and the precoding-based beamforming. Since beamforming now can bear morethan one stream and there is no essential difference between the precoding-basedbeamforming and the precoding-based spatial multiplexing, the precoding-based beam-forming and the precoding-based spatial multiplexing are called as beamforming in thisdissertation. Beamforming can promote transmission rate of wireless signal remarkably andincrease the bestrow of wireless signal. However, the performance of beamforming dependsgreatly on the channel state information at transmitter (CSIT). In time division duplex (TDD)LTE systems, base station (BS) estimates the CSIT via sounding signal sent by user, but CSIdelay and channel estimation error cause imperfect CSIT, which can degrade systemperformance. In this dissertation, the impacts of CSI delay and channel estimation error on biterror rate (BER) are first analyzed, and then sounding optimization schemes are proposed inboth time domain and frequency domain separately to minimize the impact of CSI delay andchannel estimation error on beamforming system.The main contributions of this dissertation are as follows:1）The impacts of CSI delay and channel estimation error on TDD single-user MIMO(SU-MIMO) and multi-user MIMO (MU-MIMO) beamforming systems areanalyzed in depth, and the expression is derived between system BER and CSI delayand channel estimation error, which provides the theoretical guidance for the peopleto optimize the performance of SU-MIMO beamforming systems. The obtained BER(integral) expression is useful for the design of adaptive modulation and coding scheme in the practical communications systems. Moreover, it is found by simulationthat CSI delay causes the loss of diversity gain while channel estimation error causesthe loss of array gain.2）A sounding time-domain optimization scheme, dynamic time interval scheme, isproposed to minimize the impact of CSI delay. Since the current LTE Soundingschemes can’t adapt to the change of channel environments, the dynamic timeinterval scheme is proposed which can adapt to the change of channel environments.Given the normalized pilot overhead, the optimal pilot time interval is obtained in aclosed-form expression, which depends on Doppler shift and uplink pilot SNR(signal to noise ratio). In the proposed scheme, the optimal pilot time intervalchanges as Doppler shift or/and uplink pilot SNR change(s), so the proposedscheme is robust to the change of channel environments. Moreover, because theoptimal pilot time interval is of closed-form expression and both Doppler shift anduplink pilot SNR can be measured, the proposed scheme can be implemented in apractical communication system. In addition, although the aperiodic soundingscheme is introduced in3GPP Release10and the sounding time interval can beadjusted, the formula to calculate the sounding time interval is not provided. Theclosed-form expression of the optimal pilot time interval in the proposed scheme isexactly used to calculate the sounding time interval for the aperiodic soundingscheme.3）For SU-MIMO systems, a frequency-domain sounding scheme, Frequency SelectiveSounding Reference Signal scheme (FS-SRS scheme), is proposed to mitigate theimpact of channel estimation error. In the current LTE Sounding schemes, there existsounding interferences among different users in case of fading channel, whichdegrades the accuracy of CSIT and thus beamforming performance. So FS-SRSscheme is proposed to eliminate sounding interferences. The proposed FS-SRS scheme not only obtains the better CSIT accuracy but also has some advantages overthe current LTE Sounding schemes, for example, it is robust to frequency selectivechannel, it is almost not impacted by timing offset in a practical communicationsystem, it is adaptive to the change of the underlying scheduler, and it decreases thesounding load and increases the effective sounding capacity.4）For MU-MIMO systems, an extended-FS-SRS scheme is proposed based on FS-SRSscheme. Compared to the FS-SRS scheme, the extended-FS-SRS scheme allowsmore than users to be admitted in one transmission time interval (TTI), which is veryvaluable for the enhanced downlink MIMO and Coordinate Multipoint (CoMP) inLTE-Advanced. Especially the proposed FS-SRS scheme and extended-FS-SRSscheme are applicable to the uplink sounding design in LTE-Advanced systems.