| Blind identification is an important signal processing technology, which estimates the channel status information by using the system's outputs and some auxiliary information. Due to the pursuit of high capacity and high reliability in mobile communications, the technique of channel identification with no training sequence is very attractive. The SOS (Second Order Statistics) contains not only the amplitude information of the channel, but also the phase characteristics. Therefore, the blind identification of non minimum phase system can be achieved by the introduction of cyclostationarity.Firstly, this paper introduces the basic concepts of wireless mobile channel and two equivalent SIMO (Single Input Multiple Output) channel models, then the general identifiably conditions based on the two models are given. Moreover, the CR (Cross Relation) algorithm is simulated, which explains the general identifiability conditions.Secondly, the subspace algorithm of SIMO system is presented. This method can identify the channel status information only by using SOS in which the cyclostationarity is introduced through a single sample or multiple output sampling at the output, which provides important theoretical basis for the blind identification. The traditional subspace methods require accurate channel length estimation or rank estimation of the correlation matrix and demand a lot of computation for the SVD (Singular Value Decomposition). In order to overcome the two disadvantages, the subspace algorithm on the basis of ULV updating and successive cancellation techniques is studied, and corresponding batch and adaptive algorithms are proposed.Finally, a special kind of cyclostationarity is introduced by modulation. With these two kinds of cyclostationarity, the subspace algorithm is put forward, which can not only obtain the blind identification of SISO (Single Input Single Output) system, but also improve the performance in anti noise and anti interference. |
PDF Full Text Request