Research On MIMO Performance Based On Measurement

With the rapid development of the smart terminal, multimedia service and mobile internet, more and more data traffic occurs in indoor environment, and it raises a great challenge to the future mobile communication systems to meet the peak data rate and spectrum efficiency. Increasing the system bandwidth and using high spectrum efficiency transmission techniques are very effective methods to solve the problems. Due to the current spectrum resource constrain-t, only the resource in high frequency band is available for mobile communi-cation systems. MIMO technique can greatly enhance the system spectrum efficiency without additional bandwidth and transmit power. There are rich s-cattered and reflected object in indoor environment, and the channel is usually slow time-varying, so it provides the favorable condition for the application of the MIMO technique. Hence, it is very important to study the closed-loop MIMO and the adaptive techniques based on the high frequency band channel measurement. The major work and contributions of this dissertation includes:(1) Measurement-based channel capacity analysis in indoor environ-ment:Under the assumption of fixed received SNR and fixed transmit power, the influence of the different propagation conditions on the channel capacity characteristics is studied, especially on the average channel capacity. For the fixed received SNR, the results indicate that considerable average channel ca-pacity is achieved in indoor LOS scenario, but the capacity in LOS case is more position-sensitive compared that in NLOS case. It verified the availability of MIMO technique in indoor environment. Furthermore, the influence of large scale path loss on the channel capacity is discussed.(2) Theoretical study on LOS MIMO channel capacity enhancement: Based on Wishart random matrix theory, the joint ordered eigenvalue proba-bility density function is derived for LOS MIMO systems with reconfigurable array. The moment generating function formulas of the mutual information is then obtained. The exact expression for the mean mutual information is also de-rived, and validated by simulations. The results indicate that the capacity gain of the reconfigurable array is more obvious for high order antenna array, and the potential application scenarios include indoor wireless access, short-range communication and so on.(3) Measurement-based performance analysis of codebook-based MI-MO systems:Considering closed-loop MIMO systems with quantized code-book, the exact capacity loss expressions for beamforming and full-rank mul-tiplexing transmissions are derived. Low complexity codebook selection cri-terion is obtained utilizing proposed capacity loss theory, and the criterion for full-rank multiplexing transmission can also be applied for multi-user MIMO case to reduce the inter-user interference level. Based on the MIMO channel measurement in typical indoor scenario, the relationship between the codebook capacity loss and channel condition is studied. Also the performance of differ-ent codebook selection criterions, the fluctuation level of key MIMO channel metrics and the effect of the frequency domain feedback interval are evaluat-ed in real indoor channel. The results show that current codebook designed for beamforming in LTE standard is nearly capacity lossless, but intolerable capac-ity loss is caused by the full-rank multiplexing codebook. Moreover,"capacity floor" effect is observed as the frequency domain feedback interval increases. The research is the basis for indoor MIMO enhancement evolution.(4) Research on average codebook capacity loss fitting theory:Based on the channel measurement, average capacity loss fitting theory is originally proposed. The capacity loss for beamforming transmission is mainly influ-enced by the number of of transmit antennas, which is caused by the channel quantization error. The results proved that the average capacity loss for beam-forming transmission can be fitted as a polynomial function of the minimum channel eigenvalue. The codebook dimension has a great influence on the ca-pacity loss with full-rank multiplexing transmission, and large codebook di- mension could effectively suppress the inter-stream interference to reduce the capacity loss. For full-rank multiplexing transmission using LTE2×2code-book, capacity loss "cutoff" phenomenon is captured. In other words, there exists a looser capacity loss lower bound for current LTE2×2codebook, and it should be avoided in future system design to the greatest extent. The av-erage capacity loss for full-rank multiplexing transmission can be fitted as a power function of the minimum channel eigenvalue. Due to the random code-book quantization error, it is difficult to directly study the channel capacity for codebook-based MIMO systems. Fortunately, the problem can be easily solved utilizing proposed average capacity loss fitting theory, and simple channel er-godic capacity expressions for beamforming and full-rank multiplexing trans-missions are obtained in the high SNR region. The proposed theory initiates a novel method to study the channel capacity for closed-loop MIMO systems.(5) Measurement-based framework for adaptive transmission mode selection in codebook-based MIMO systems:Considering N×2codebook-based MIMO systems, the mathematic relationship between the transmission mode selection criterion and channel condition is originally proposed. The transmission mode selection criterion is expressed as a polynomial function of the SNR and minimum eigenvalue, and the mode selection relies on the sign of the decision function. Moreover, the polynomial function can be easily real-ized in hardware platform (e.g. DSP) through simple multiply-add operations. Utilizing the proposed average capacity loss fitting theory, high SNR analysis on the transmission mode selection is given. It is proved that the proposed transmission mode selection criterion is equivalent to maximize the channel er-godic capacity. Finally, the simulation results show that the proposed adaptive transmission mode selection criterion is nearly capacity lossless, and the gen-erality is validated using channel measurement and channel models in different frequency band, scenarios and antenna polarized methods. Moreover, the pro-posed framework can be generalized for other antenna configurations and new codebook design standardized in future mobile communication systems.