Research On Signal Detection Algorithms For Spatial Modulation Systems

The rapid development of wireless communications leads to the ever-increasing demand of ubiquitous wireless access,higher bitrate,better spectrum efficiency and lower implementation complexity.Multiple input multiple output(MIMO)technology can achieve broadband communications with higher spectrum efficiency.However,the conventional MIMO technology faces several technical obstacles,including inter-channel interference(ICI),inter-antenna synchronization(IAS),multiple radio frequency chains,large power consumption and high demodulation complexity due to ICI.Specially,these disadvantages turn to be more serious in the large-scale antennas systems(LSAS).To overcome these shortcomings of conventional MIMO systems,spatial modulation(SM)technology was proposed as a novel multiple antenna transmission scheme.SM has been a hot topic in the field of current wireless communications,and it holds promising potential for the next-generation large-scale wireless communications.To overcome the problem of high computational complexity of the optimal maximum likelihood(ML)detection algorithm,we intensively studied the signal detection problem at the receiver in the SM and generalized space shift keying(GSSK)systems,respectively.The main works include:(1)We studied the existing detection algorithms of the SM systems,and emphatically studied the M-algorithm to maximum likelihood(MML)detection algorithm.In MML algorithm,the number of reserved nodes per layer in the search tree can directly affect the bit-error-rate(BER)performance and the computational complexity,this work adopts two kinds of selection methods.One is the fixed M-value selection method,in which a simple geometric progression rule is used to reserved nodes from the first layer to the last layer.Another is the dynamic M-value selection method,which is based on the adaptive nodes reservation.In addition,since the detection order of MML detector is restricted to the ascending order of the receiver antenna indices which neglects its huge impact on the BER performance,we put forward several enhanced MML algorithms based on the status of channel conditions,the receive signal strength,and both of them.The BER performance comparisons of the proposed algorithms with existing detection algorithms show the advantages and disadvantages of the proposed algorithms,respectively.The detection performance and the computational complexity are then presented from two aspects of theory and simulations.The simulation results show that the proposed algorithms outperform the MML algorithm as well as has the same computational complexity.(2)We deeply studied the existing detection algorithms for the GSSK systems,aiming at the binary property of the each element of the transmitted symbol vector,the detection of GSSK signal can be converted to binary quadratic programming problem.A new GSSK detection algorithm based on the global optimality conditions for binary quadratic programming is proposed.The BER performance and the computational complexity comparison of the proposed algorithm with the existing detection algorithms are given.The detection probability of each element of the transmitted signal vector using optimal decision rule is confirmed from the theory and simulation.The simulation results show that the proposed algorithm can reduce the computational complexity as well as maintaining a near-optimum BER performance.(3)The detection of GSSK signal based on sparse reconstruction(SR)theory is studied.When the number of activated antennas is far less than the number of transmit antennas,the transmitted signal vector is sparse.Based on the sparse reconstruction theory,we propose two detection algorithms,SR and iterative SR(ISR)algorithms.Utilizing a non-convex arctangent penalty function,the proposed SR algorithm can obtain more sparse result than e1 norm regularization.The detection performance is superior to other sparse reconstruction based algorithms.In order to further improve the detection performance of SR algorithm,we proposed ISR algorithm wherein SR algorithm is applied through several iterations.From the BER performance and the computational complexity perspective,we compare the proposed algorithms with the related detection algorithms.The simulation results show that the BER performance of the proposed algorithms outperforms the existing detection sparse reconstruction based algorithms.