Study On Signal Blind Processing Of MIMO And MIMO-OFDM Systems

With the rapid development of mobile communications,spectrum resources have become increasingly tight,and users have more urgent requirements for the transmission rate and channel capacity of communication systems.The multiple input multiple output(MIMO)communication system utilizes spatial multiplexing and spatial diversity technologies,which can significantly increase the transmission rate and channel capacity of the system without adding additional bandwidth.Orthogonal frequency division multiplexing(OFDM)communication system rely on mutually orthogonal subcarriers to transmit information,which not only resists the effects of frequency selective fading channels,but also improves the efficiency of the spectrum.A MIMO-OFDM system combining MIMO and OFDM can have the advantages of both.With the increasing application of MIMO and MIMO-OFDM systems,it is becoming increasingly important to study the blind signal processing of these two systems in non-cooperative communication scenarios.This thesis focuses on signal identification and parameter estimation in MIMO and MIMO-OFDM systems.The main work is as follows:1.Aiming at the blind identification of orthogonal space-time block codes(OSTBC)in MIMO system,a recognition method based on joint approximate diagonalization of eigenvalue matrix(JADE)and feature extraction is studied.Using the idea of blind source separation(BSS),the received MIMO signal is first converted into a linear instantaneous hybrid model in BSS,then the sources number estimation and whitening of the model are done,and then the JADE algorithm is used to estimate the virtual channel matrix of the model.According to the orthogonality of the coding matrix of OSTBC,it can be concluded that the correlation matrix of the virtual channel matrix of OSTBC is a quantity matrix.Finally the blind recognition of OSTBC is realized by extracting the feature parameters from the quantity matrix.2.Aiming at the blind identification of MIMO system modulation,a recognition method based on JADE and feature extraction is studied.According to the independence of the signals on each transmitting antenna of the spatial multiplexing MIMO system,the number of transmitting antennas is estimated and the whitening process is performed on the received signals.Then the transmitted signals are separated from the whitened signals by using JADE algorithm,and then four characteristic parameters are extracted from the transmitted signals by using sixth-order cumulant,cyclic spectrum and fourth-power spectrum algorithm,finally these four characteristic parameters are input into the hierarchical back propagation(BP)neural network to identify the modulation mode of the signal.The simulation results show that the method can identify eight modulation modes {2PSK,2ASK,2FSK,4PSK,4ASK,MSK,8PSK,16QAM} in MIMO system.3.Aiming at the blind estimation of MIMO-OFDM signal parameters,a multi-parameter estimation method based on cyclic autocorrelation and fourth-order cyclic cumulant is studied.First,the autocorrelation function of MIMO-OFDM is calculated,and the cyclic autocorrelation function is obtained after Fourier transform.The symbol period,useful symbol period and cyclic prefix length of the signal can be estimated by analyzing the different sections of the cyclic autocorrelation three-dimensional graph.Then,the fourth-order cyclic cumulant of MIMO-OFDM is derived,and it is found that when the cyclic frequency is taken as the subcarrier frequency,a peak of the fourth-order cyclic cumulant will appear.The number of subcarriers and subcarrier frequencies of MIMO-OFDM can be estimated by detecting the number of fourth-order cyclic cumulant peaks and their corresponding cyclic frequencies.