Mimo Detecting Research And Implementation Based On General Purpose Processor

LTE and LTE-A are highly valued by developers in recent years, more and more researchers make their contributions to the key technique of the4G communication system. MIMO detecting, as the crucial technology to expand the throughput of the wireless system, is one of the most important one among all the technologies of LTE/LTE-A system. The improvement of MIMO technique will benefit the4G system.Meanwhile, the appearance of Software Radio platform SORA presented by Microsoft in2008attracted the attention of the developers in signal processing area. The developers find huge potential for General Purpose Processors (GPP) and are making constant efforts to implement real-time signal processing system on the GPP platform.In this environment, this paper focuses on the research and implementation of downlink MIMO detecting on general purpose processor platform, one of the crucial point to realize real-time4G signal processing system on GPP. The research scope includes MIMO algorithm research, high effective implementation methods of applying MIMO detecting on GPP and a method of using SINR estimation to improve the performance of the MIMO system.In the first place, we analyze the LTE/LTE-A system and the MIMO technology in it. With a traditional implementation of several MIMO detecting algorithms, we compared the performance and difficulty of the algorithms and choose one of them to apply high effective methods on. Secondly, with the optimization implementation of MIMO detecting, we estimate the performance of the realization of real-time signal processing system on GPP, we also apply multicore technology to further accelerate the processing and make some experiment about it. At last, in order to improve the throughput of the whole system, we complete some extra research on how to use SINR estimation of MIMO to enhance the bit error performance without bring much real-time loss. A large amount of experiments have been completed to confirm the best implementation configuration on GPP platform. This research helps the system reach the requirement of the throughput in protocols which we will exhibit at the last of the paper.Our results show that, the MIMO implementation of GPP platform cannot reach the real-time requirement of protocols at present, however, the multicore technology and the cloud computing gave GPP platform large potential in the future development. The cloud computing will provide plenty of computing resource in GPP mode, and the combination of such resource and the large amount of computing task generated by signal processing of next generation communication system will lead to a effective use of resources, which can be realized by implementing signal processing systems on GPP platform. It may be a new trend of wireless signal processing system realization in the future.