Separation Method Of High-order MPSK Signals In Single Channel

The separation of single channel mixed signals is one of the important topic in the field of signal processing.In contrast with the multi antenna systems,the single antenna systems have the advantages of low cost,low energy consumption and simple system design.On the other hand,with the rapid development of heterogeneous communication network,the phenomenon that two or more signals occur in a single channel is becoming more and more common because of the spectrum reuse.However,since the single channel mixed signal is partially or completely overlapped in time domain and frequency domain,it is impossible to separate the mixed signal by using the traditional time domain or frequency domain filtering method.Therefore,it is of great practical significance to study the separation method of multi-component mixed signals received by the single antenna system.This thesis focuses on the separation of high order MPSK digital modulation signals in a single channel receiver.On the basis of constructing the virtual multi-channel model,the signal separation method based on semi-definite relaxation and low rank semi-definite relaxation is investigated.The main work of this thesis are as follows.1)A signal separation method based on semi-definite relaxation is proposed.Firstly,the virtual multi-input multi-output(MIMO)model is constructed by over-sampling the received signal,and the discrete constraints for input signal variables is relaxed to a semi-definite constraints.Consequently,a semi-definite programming problem is established.Secondly,by utilizing the interior point algorithm and the randomized method,the semi-definite programming problem is resolved,and the input symbol sequences is estimated.Finally,considering the special structure of the virtual MIMO model,the decision feedback equalization is introduced to improve the signal separation performance.2)A signal separation method based on low rank semi-definite relaxation is presented.First of all,a rank-1 constraint is added to the semi-definite programming problem,thus a low rank semi-definite programming problem is formulated.And then,the alternative variable algorithm is used to optimize this problem,which can transform the multi-variable optimization into the single variable optimization.In the second place,some strategies and methods are fully discussed to obtain the near optimal solution of original problem,such as the selection of starting point,the disturbance of local optimal solution,symbol decision and so on.At last,the feedback mechanism is also introduced to obtain the improvement of separation performance.