| With the explosion of communication application market and the increasing number of users, it has become a long-term hot topic for wireless communication technology research to achieve more effective communication while alleviating the scarcity of spectrum resource. Millimeter wave communication has attracted much attention from academia and industry, because it owns large bandwidth and enables high-speed data transmission achievable with excellent anti-interference performance, etc. Due to the limitation of implementation, the non-ideality of RF analog front-end inevitably occurs, which can degrade the communication system performance. Therefore, it is very necessary to take the distortions into account for more accurate performance evaluation and to compensate them with digital processing.This thesis mainly investigates two typical nonlinear distortions in RF analog front-end: IQ imbalance(IQI) and phase noise(PN). Here, the system model, the channel model and frame structure of transmission following the IEEE 802.11 ad standard will be adopted for the impact analysis and compensation scheme research for the two nonlinear distortions. Single Carrier Frequency Domain Equalization(SC- FDE) will be chosen since it has lower Peak to Average Power Ratio(PAPR) than Orthogonal Frequency Division Multiplexing(OFDM)The main content of the thesis can be roughly divided into three parts, and the general arrangement is as follows:The first part introduces the research background of 60 GHz communication, which is typical in the millimeter wave communication, then its system design and main parameters following the IEEE 802.11 ad standard would be presented. Besides, this part presents a brief overview of the two main nonlinear distortion mentioned above.The second part focuses on phase noise. Firstly, the phase noise model compliant with IEEE 802.11 ad standard, the system model and the impact analysis will be presented. Then, the research and analysis of the Common Phase Error(CPE) compensation scheme, which is commonly used, will be carried out. Considering the limited compensation of CPE, this thesis proposes a new compensation scheme, which reconstructs the phase noise through the eigenvalue decomposition for phase noise autocorrelation matrix in time domain. Further, to obtain the more accurate estimation of phase noise for compensation, the Sparse Bayesian Learning(SBL) is tailored here to estimate the coefficients for reconstruction. Through the the theoretical derivation and simulations, it verifies that the proposed compensation scheme can achieve an obvious improvement for system performance.The third part focuses on IQ imbalance. Generally, the IQI model can be classified into two types, one is the frequency-independent IQI caused by the inaccuracy of local oscillator signal, the other is the frequency-dependent IQI which takes the difference between the frequency responses of I branch and Q branch into account. As for the mentioned two situations, this part respectively propose a training-based compensation scheme. For the former, the thesis presents a compensation scheme in which the Least Squares(LS) estimation for IQI is based on the minimization of the Channel Variation Energy(CVE). For the latter which is more complex, a compensation scheme based on Maximum Likelihood(ML) iterative estimation is proposed. The theoretical analysis and simulations show that both of the proposed compensation schemes have achieved very effective compensation performance. Besides, compared with other existing IQI compensation schemes, they can obtain IQ imbalance parameters separating from channel estimation. What’s more, two proposed schemes have no specific requirements on the training sequence, so that they won’t bring in any additional data or the change of the frame structure, which lead to a wide application scope. |
PDF Full Text Request