Research Of Linear Multiuser Detect In CDMA Communication System

In CDMA mobile communication system, because of the multi-user's connection, the spread spectrum codes they use are normally not quadrature rigidly. The nonzero cross correlation coefficient between codes will arouse the interference between the users, which is called multiple access interference. Because the spread spectrum code of the users is known, the cross correlation coefficient between the users is known. The receiver can get some important information from multiple access interference,and can detect associate or eliminate the mutual interfere from the receiving signal. The strategy that utilizes many users information while detecting is called multi-user detect. It is the method that can restrain multiple access interference, which is approbatory by research personnel. The optimal multi-user detector that has the optimal capability indexes is composed of matching filter and Viterbi coding algorithm in theory, but because the operation complexity is increased with the user exponent(),it can't be actualized in fact. The hypo-optimal multi-user detector becomes the perfect choice, and becomes the hotspot in research recently.Generally, in CDMA mobile communication system, the sink's receiving signals are composed of multi-user's useful information and channel noise. There are inter- symbol interference and multiple access interference, and the basis of channel noise is Gauss white noise. The channel noise will be regarded as Gauss white noise totally in this article. The main purpose of multi-user detector is to eliminate multiple access interference. The hypo-optimal multi-user detector that relatively integrity in theory and proximal to optimal multi-user detector is discussed in this article, and Decorrelating Detector, Minimum?Mean?Square?Error Detector, Optimal linear Detector together with the multi-capability indexes (output energy, SNR, bit error rate, asymptotic efficiency, etc.) are studied and analyzed in this article. On the base of deep research of linearity multi-user detector, a new blind multi-user detector is put forward----minimum output energy combined with subspace detector. It can restrain multiple access interference, and the same time eliminate Gauss white noise of Gauss feature, namely channel noise.In this blind multi-user detector, the signal parameters must be known are characteristic wave and time information of anticipant signal. While the traditional multi-user detector needs more information (including characteristic wave of interferential signal, time information and relative amplitude), though it need one more condition of anticipant signal's time information than self-adapting multi-user detector, it needs no training sequence, and decrease operation complexity, shorten time delay, make real time detection come true.In the analytic approach of signal subspace, input signal can be decomposed according to eigenvalue, and get a decomposing matrix composed of signal subspace and noise subspace, the formula are as below: In numeration, correlation matrix composed of noise subspace can be taken out directly, then the main part of channel noise of Gauss feature is eliminated, and get the correlation matrix , composed of signal subspace which is made up of estimate signal. As the theory about decompose of subspace said above, a signal that interfered by noise, in fact is the overlapped vector of signal and noise. Seen from space, we can recognize the space the useful signal lies in is signal space, the space the noise lies in is noise space. Project the polluted signal on signal subspace and noise subspace separately, substituting the projection signal component in signal subspace for the original signal to calculate can eliminate the interference of noise. The detecting capability is better than detecting in original signal space. According to minimum output energy rule, we can take minimum output energy multi-user detector (MOE detector). In MOE detector algorithm, we adopt RLS, substitute correlation matrixcomposed of si...