The Application Of The 2-Order Cyclostationarity Of Signals In Wireless Mobile Communication

The dissertation focuses on the blind signal processing in wireless mobile communications based on the cyclostationarity of signals. In particular, blind estimation of signal-to-noise ratio and synchronization parameters of OFDM systems based on cyclostationarity and their application in time-varing channel are thoroughly investigated in this dissertation.SNR is an important parameter in wireless communication, not only as the indicator of channel quality but also as the necessary input information for implementation of some algorithms. A blind SNR estimation based on the cyclostationarity of received signals is proposed in this dissertation. Firstly, the second-order cyclic autocorrelations of received signals on the non-zero cyclic frequencies are used to estimate the power of signal. Then, noise power is estimated according to the relationship between the 2nd-order moment of the received signal and the power of signal and noise. Then estimated SNR can be calculated easily. This estimation can be applied in MPSK or MQAM modulation, real or complex channel, one-path channel or multipath channel, and performs better than other classic blind SNR estimations, such as M2M4 and SVR estimation in a wide SNR region. Also, its normalized MSE is closer to a Cramer-Rao lower bound.As a multicarrier system, OFDM is far more sensitive to synchronization error than the other single carrier systems. So synchronization has been one of the major research topics in OFDM system. Firstly, the OFDM system is discussed in which one user is allocated one subcarrier. The dissertation introduces an improved blind synchronization parameters estimation based on Bolcskei's estimation. Also for conjugate cyclostationary signal, a modified frequency offset estimation exploiting the second-order conjugate cyclic autocorrelation is proposed. Then, the OFDM system in which one user is allocated more than one subcarrier is discussed. For the general subcarrier allocation schemes, the synchronization parameters estimation is proposed. In this estimation, the estimation of frequency offset is relying on the estimation of time delay, which would degrade the performance. Then aiming at the three significant subcarrier allocation schemes, including block allocation, interleaved allocation and hybrid allocation, a new frequency offset estimation independent of the time delay estimation is proposed. This estimation of frequency offset has higher resolution. At last, the OFDM system in which all subcarriers bear data synchronously is investigated.The dissertation discusses in detail the application of signal processing based on the cyclostationarity of signals to time-varying channel. Firstly, computational complexity of estimation of cyclic statistics is investigated. The computational complexities of three methods, including conventional estimation, real time updated estimation and real time updated estimation based on FFT, are compared. Then, time-varying channel is approximated and simplified to the model in which amplitude and phase are both linearly changed. In this simplified model, the cyclostationarity of received signal can be proved. Also the estimation of the cyclic statistics of received signal is introduced, and is proved in three different simulation scenarios. Then, the relation between cycle number in the coherence time and the veracity of the cyclic statistics estimation is showed, and the experiential formula to select cycle number is presented. According...