Research On Receiving Technology Based On Compressive Sensing In MUSA

One of the main objectives of the fifth generation of mobile communications is to support massive Machine Type Communication(m MTC).Non-orthogonal Multiple Access can provide more multiplexing resources to meet the needs of mass machine access.Multi-User Shared Access(MUSA)technology itself is a grant-free transmission mechanism,which has the advantage of low signaling overhead in the case of m MTC,and has attracted much attention.Considering the users in grant-free MUSA transmit data sporadic bursts,the receiver should detect the user behavior before it can detect data for active users,and the sparse signal recovery in the compressive sensing theory can simultaneously detect the user's active behavior and data information.In this thesis,the receiving technology of MUSA system is studied in combination with compressive sensing,and the innovations are as follows:1.The Validation Error-Orthogonal Matching Pursuit(VE-OMP)algorithm is proposed to break the limitation of known sparsity in OMP algorithm and the limitation of known noise power as the stop threshold in Sparsity Adaptive Matching Pursuit(SAMP)algorithm.When the symbol error rate reaches the 10^-3,VE-OMP algorithm improves 2 d B compared with OMP algorithm in performance.VE-OMP algorithm can end iteration adaptively and more practical than OMP algorithm;VE-OMP algorithm is inferior to SAMP algorithm in performance,but it has lower complexity than SAMP algorithm,and the judgment of the end of iteration is not affected by the noise environment,which is more flexible.2.Aiming at the problem of fast channel information failure in the scenario of mass equipment grant-free access,the joint channel estimation and multi-user detection are proposed,and the model is established.Two block sparse matching pursuit algorithms are given for the joint channel estimation and multi-user detection of uplink MUSA.The simulation results show that the sensing success rate of joint detection reaches 1 when the signal-to-noise ratio(SNR)is 4,that is,perfect perception.The symbol error rate performance of joint detection is about 4 d B better than that of independent detection at low SNR,and even better at high SNR.Joint channel estimation and multiuser detection greatly improve the performance of the receiver.3.The spreading sequences of MUSA ensure low cross-correlation for high detection performance.When the number of users is huge and the overload rate is fixed,the length of sequence increases,so it is difficult to guarantee low cross-correlation.In this thesis,spreading sequence group is proposed as the multiple access signature instead of the single sequence spreading.The principle of the transmitter is unchanged,which can be regarded as the whole spreading of the symbols in the group.When the SNR is greater than 2 d B,the sensing success rate of the sequence group spread spectrum scheme reaches 1,and the errors performance of the receiver can approach the theoretical lower bound of the least squares algorithm with known support set.The spreading sequence group scheme improves the sensing success rate of the receiver without reducing the overload rate.