Research On Key Technologies Of Blind Source Separation In Wireless Communications

In recent years, blind source separation(BSS) has been widely applied in many areas for its outstanding technology advantage, which has become a research hotspot in the field of signal processing. Blind source separation technology can relax the restricted condition of a priori information in wireless communication systems. It can recover the unobserved unknown sources only from the received mixed signals based on features of the source signals. It is an important theory to realize high spectrum efficiency, strong anti-interference and adaptive signal processing in wireless communications. In order to optimize and enhance the performance of the wireless communications which includes improving the system spectrum efficiency, anti-interference ability and the signal detection performance, and to satisfy the development requirement of future wireless communications, this dissertation investigates the key technologies of blind source separation in wireless communications to achieve the goal of blind adaptive receiver signal processing of wireless communication systems. The work can be mainly divided into the research of blind source separation algorithm and the key technologies of blind source separation in wireless communication systems.The main content includes the following:For the noise sensitive issues of blind separation algorithm, combining with the bit error rate(BER) performance metrics of communication systems, a blind source separation algorithm is proposed based on the minimum bit error rate criterion. The principle of the proposed algorithm is: firstly, the new minimum bit error rate criterion constrained cost function is constructed according to the derived minimum bit error rate criterion incorporated into maximum likelihood(ML) principle; secondly, by virtue of nature gradient descent search, the blind separation is accomplished through minimizing this cost function. Simulation experiments and analysis show that, in comparison with the blind separation algorithm based on the cost function of the maximum likelihood principle, the proposed algorithm based on minimum BER criteria constrained cost function can lead to better performance in speed of convergence and separation accuracy.In view of the estimation problem of mixing matrix of the underdetermined blind source separation(UBSS) in communication scenarios, a novel underdetermined blind identification algorithm is proposed. This proposed algorithm employs the statistical and structure property of generalized covariance and the compressive characteristic of Tucker decomposition. Firstly, the core functions are built based on generalized covariance matrices. Then the core functions are stacked as a three-order tensor, and the tucker decomposition of constructed tensor is executed to estimate the mixing matrix. The proposed algorithm not only has the better identification performance, but also the lower computational complexity. The simulation experiments demonstrate the effectiveness of the proposed algorithm.The problems of blind multiuser detection and blind code estimation in direct sequence-code division multiple access(DS-CDMA) systems are investigated. For the computational complexity problem of blind separation algorithm due to higher order statistics, a blind adaptive algorithm is proposed based on generalized covariance matrix, which is used to carry out blind separation of user signals and blind estimation of spreading sequences in DS-CDMA systems. This blind separation algorithm exploits charrelation matrix for its simple computation and effective extraction of information from observation signal samples to improve separation performance. The system model of DS-CDMA signals is modeled as a blind separation framework. The unknown user information and spreading sequence of DS-CDMA systems can be estimated only from the sampled observation signals. Theoretical analysis and simulation results show that the proposed algorithm can effectively realize blind user separation and blind code estimation for DS-CDMA systems when the number of observation samples is less and the signal to noise ratio(SNR) is low.A second-order cone(SOC) constraint based robust blind multiuser detector is proposed for DS-CDMA systems against signature waveform mismatch(SWM) problem derived from the influences of time asynchronization and channel fading. Firstly, DS-CDMA systems with SWM problem is formulated as blind source separation(BSS) model subject to the second-order cone constraint. Secondly, the resulting blind separation based on the second-order cone constraint problem is solved by approximate negentropy maximization using quasi-Newton iterative methods to realize the blind multiuser detection. Theoretical analysis and simulation results show that the separation performance of the proposed blind detector is superior to those of the existing methods, and has the advantages of low complexity.The problems of interference elimination and sources recovery in OFDM systems are investigated. A blind adaptive interference suppression scheme based on blind source separation is developed to overcome the inter-carrier interference(ICI) of orthogonal frequency division multiplexing(OFDM) systems subject to unknown carrier frequency offset(CFO) and multipath. Taking into account statistical independence of subcarriers’ signals of OFDM, the signal recovery mechanism is investigated to achieve the goal of blind equalization. The received OFDM signals can be considered as the mixed observation signals. The effect of CFO and fading channel corresponds to the mixing matrix in the problem of blind source separation framework. The ICA-based blind separation of the OFDM system model is built, and the cost function of ICA-based detector is obtained based on minimum mutual information principle. The natural gradient is used to optimize the cost function and extract source signals from the observation of a received mixture for the carrier synchronization. The blind separation technique can increase spectral efficiency and provide robust performance against the erroneous parameter estimation problem. Theoretical analysis and simulation results show that compared with the conventional pilot-based scheme, the performance of OFDM systems is improved by the proposed ICA-based blind detection technique.In view of tensor decomposition based blind separation OFDM model, a Vandermonde constrained tensor decomposition based blind separation approach is investigated to overcome ICI of OFDM transmission subject to unknown CFO and fading channel. The main idea is that signal reception model is formulated as a tensor decomposition framework involving a Vandermonde structure factor matrix. Then an efficient linear algebra algorithm is resorted to the computation of the factors for ICI suppression and symbol recovery in OFDM systems. Studies have shown that the proposed algorithm takes the Vandermonde structure into account in tensor model, which not only improves the performance of the interference elimination and symbol recovery of OFDM systems, but also reduces the computational cost compared with the existing unconstrained tensor decomposition method. The simulation experiments verify the effectiveness of the proposed algorithm.