Research On Spherical Decoding For Generalised Spatial Modulation

As a new multi-antenna technology,Spatial Modulation(SM)technology,selectively activating a single transmitting antenna,uses the number of the working antenna and modulation symbol to collectively transmit the information,and overcomes the increase of complexity and cost of the MIMO systems,caused by inter-channel interference,inter-antenna synchronization,multi-radio link and other reasons.Generalized spatial modulation(GSM)technology is the development of spatial modulation technology,which activates multiple transmit antennas at the same time.Compared with the SM systems,GSM systems can further improve the system throughput and t he utilization of the transmitter antenna combination.In the signal detection algorithms of the spatial modulation systems,the optimal algorithm of error performance is the Maximum-Likelihood(ML)detection algorithm.but its computational complexity increases proportionally with the number of transmit and receive antennas and the size of the constellation.So it’s difficult to apply it to the practice systems.While the application of Sphere Decoding(SD)algorithm to spatial modulation system can signif icantly reduce the computational complexity of the signal detection.In this paper,three kinds of spatial modulation-spherical decoding(SM-SD)algorithms are extended to the GSM system.They are the Receiver-centric spherical decoding algorithm(Rx-SD),the transmitter-centric spherical decoding algorithm(Tx-SD)and the Combined spherical decoding algorithm(C-SD),which is a combination of Tx-SD and Rx-SD.The bit error rate and computational complexity of the three algorithms are analyzed.It is proved that by selecting a reasonable spherical decoding initial search radius,the G SM-SD algorithm can achieve the same BER as the GSM-ML optimal detection,and can significantly reduce the computational complexity.According to the theoretical analysis and simulation,none of the three algorithms is always optimal.The choice of the optimal algorithm is related to the number of transmitting and receiving antennas,the modulation mode and the signal-to-noise ratio of the GSM system.The scope of application of several algorithms are given in this paper.In addition,we apply a new channel realization method to the GSM-SD algorithm,and propose an improved spherical decoding algorithm.The code search tree of the modified GSM-SD algorithm is novel and simplifies the search flow,and the new channel realization approach will bring about the benefits of reducing the computational complexity.Finally,the simulation results show that the modified algorithm has the same performance as the GSM-ML detection algorithm,and the computational complexity is the lowest compared with other existing GSM-SD algorithms at low SNR.