Research On Mobile Communication Systems With Imperfect Channel State Information

System research and analysis with imperfect channel state information (CSI) play a very important role in communication area. The paper aims to study and analyze system performances on condition of imperfect CSI. Characteristics of mobile radio channels, estimation methods and algorithms, comparisons between perfect conditions and imperfect ones and so on are main contents in this paper. Especially, pilot symbol assisted modulation (PSAM) scheme is a main topic here. Interest questions include how to design pilot symbols, such as the pilot-insertion frequency and the power arrangement between pilot symbols and information symbols, to guarantee the minimum channel estimation error and the required quality of service (QoS). The function and impacts of channel estimation on future mobile communication systems with advanced techniques have also been investigated here.Moreover, the work focuses on the acquirement of channel information. The minimum mean square error (MMSE) algorithm is used for channel estimation and prediction. PSAM is viewed as a facilitated method to acquire channel information. PSAM is an efficient channel tracking scheme and has been adopted by many communication system standards, such as WCDMA, CDMA2000 and so on. However, there are some problems caused by PSAM. Pilot symbols consume part of the transmitted power and bandwidth, which in turn reduces the spectral efficiency. So, it is necessary to design the pilot symbols optimally because it impacts both estimation error and system performance. And another necessary point is to understand the characteristics of mobile radio channels because they are related to PSAM.Systems referred here include single input and single output (SISO) system, multiple inputs and single output (MISO) system, multiple inputs and multiple outputs (MIMO) system and so on. Main techniques include orthogonal frequency division multiplexity (OFDM), adaptive modulation and coding (AMC), cross-layer design and so on. The mathematic tools include statistic and probability, the optimization and so on.Main work and innovations contain following points.1 The space correlation has been analyzed in different propagation conditions including the MIMO macro-cell and micro-cell applications. The characteristics curves describing the relationship between space correlation and antenna distance have been obtained according to the numerical analysis. Effects of the spread of angle of arrival on space correlations have also been investigated. Besides, the impact of space correlations on system performances has been discussed here. Thresholds of space correlation coefficients which influence BER performance under different propagations have also been found through BER analysis.2 In OFDM systems, the wavelet de-noising has been combined with usual channel estimation methods. The new scheme was used to estimate the fading channel gains in frequency domain. It reduced the impact of additive noise efficiently and improved system performances without extral complexity. It has also been proved that this scheme is adaptive to different channel conditions.3 Based on advantages of learning and memory in neural network, and combined with LME algorithm, a new algorithm of channel tracking as well as a novel channel estimation method has been proposed in OFDM systems. Pilot symbols were placed at a certain interval. The neural network channel estimator (NNCE) estimated both amplitude and angle values of a complex random process, accurately. And it updated values according to new pilot symbols to complete the tracking.4 Relationships between mean square error (MSE) estimation and a general power spectral of channels have been discussed. A modified expression of MSE function has been deduced. This expression includes several important channel parameters, some of which affect the estimate performance significantly, such as Rice factor, the spread of angle of arrival (AOA) and so on. Then, the performance criteria associated with estimate error function was analyzed. And impacts of these channel parameters on channel estimation and system performances were explored. Thresholds of channel signal-to-noise ratio (CSNR) in AMC systems are also influenced by MSE. This affection has also been investigated here under different propagation conditions.5 Schemes of pilot symbols in adaptive pilot symbols assisted coded modulation systems have been optimized. A new algorithm has been explored to determine the interval of pilot symbols according to the second order statistics of fading channels. Both partial parameter and global parameter were defined. The adaptation of partial parameter and the efficiency of global paremeter have been proved. It was also found that the global parameter was a tighter bound. Influences of this new algorithm on whole systems have also been tested.6 Adaptive PSAM schemes with imperfect CSI have been proposed based on cross-layer design which combines physical layer with data link layer. The purpose is to optimize system by adjusting the resource and power to the variation of channel conditions in order to meet system performances. Adaptive schemes include adaptive spacing allocation of pilot symbols and adaptive power arrangement between pilot symbols and information symbols. These schemes were employed to facilitate channel estimation and prediction. Results about impacts of these schemes on system performances as well as comparisons with perfect conditions have also been shown here. Besides, PER and ASE functions with perfect CSI have been modified considering with imperfect CSI.