| Summary: | 碩士 === 元智大學 === 電機工程研究所 === 88 === In this thesis, we consider the design and performance analysis of TDMA MMSE DFE receiver with spatial diversity, especially for the case of TDMA multiuser detection by using smart antennas.
To tackle the problem, we first consider the most fundamental case of a single-user single antenna case, i.e., the SU-SISO case. The SISO-MMSE-DFE receiver and its analytic BER performance are derived by using the Gaussian approximation technique, which treats the ISI terms as a Gaussian random variable. Then the result is further extended to single-user multiple antenna case, i.e., the SU-SIMO case. Next, we focus on the problem of TDMA multiuser detection, which means that the same time slot is allocated to two simultaneous users within the same cell. At the basestation, both the ISI and multiple access interference (MAI) should be suppressed, so we use multiple receiving antennas and a multiple-input multiple output (MIMO) MMSE-DFE receiver to detect the two users at the same time. As above, the receiver''s BER performance is derived in terms of the known channel response and some system parameters. Moreover, it is shown that if the receiving power levels of the two users are quite imbalanced due to lack of power control mechanism in TDMA existing TDMA system, the receiver suffers no near-far effect and can still achieve satisfactory performance by adjust the two decision delays to their optimum values. Interesting yet, the weaker signal can even be beneficial by the stronger signal. To confirm our theoretical works, computer simulations are conducted which show that our analytic BER prediction is quite matched to the simulation results.
It is well known that as the antenna spacing is not large enough, the fading processes for different antennas are then not uncorrelated but correlated. So the second main theme of this thesis is to consider the effects of spatial fading correlation and Doppler spread on the receiver''s performance. We first elaborate on the simulation model and procedure for generating spatially correlated time-varying fading processes. Then from the simulational results, it is noted that the Doppler rate has a much significant adverse effect on the system performance than the spatial correlation does. In fact, as long as the spatial correlation coefficient is kept below 0.5, only slight performance degradation is observed.
|
|---|
