Design And Optimization Of MMSE Based Channel Estimation Algorithm In Software Difined Radio System

Wireless communication systems have been evolved into 4G systems(LTE/LTE-A),which is still evolving towards 5 G.The future wireless systems are designed to satisfy a higher user experience and more complicated scenario.Thus,more and more cutting-edge wireless technologies are developed in 5G systems,such as new access scheme,massive MIMO,Ultra Dense Network,etc.Compared with traditional wireless systems,the peak data rate,connectivity density,end-to-end latency,user experience data rate and mobility all become the Key Performance Index in 5G systems.However,it is impossible for 5G to satisfy all the requirements using only one or few technologies,to cover all the future scenario.Thus,the future network are going to be a heterogeneous network,which is employed in an ultra-dense way.This brings the development and design of wireless systems into a hard situation,since all systems are designed in hardware way.However,Software Defined Radio(SDR)introduces a promising way to solve above challenges.SDRs are featured by high flexibility and scalability,which can realize smooth evolvement of wireless systems.What's more,heterogeneous networks make the wireless environment more complicated.And 5G are extended to higher frequency to realize higher data rate.So,the channel estimation will play more and more important role in wireless systems design.Based on above situation,this thesis is focused on the channel estimation in SDR systems,especially MMSE based algorithm.The mainly works are:1.Introduce the main features of wireless channels,and carry out simulation based on the standard wireless channel model in 3 GPP standard,which compare the advantages and disadvantages of different channel estimation methods;2.Implement 3GPP Physical Uplink Shared Channel(PUSCH)link,mainly design and implement the MMSE channel estimation module,then demonstrate the performance of RX.3.Realize the optimization of MMSE channel estimation based on GPP technologies,such as multi-core/multi-thread technology,parallel pipeline technology,SIMD processing,and high-capacity cache of GPP.By these methods,the processing efficiency of MMSE channel estimation is improved,as well as the processing delay reduced.It can be demonstrated that MMSE channel estimation can be implemented in SDR platforms,with limited increase in complexity.And by this way,it can improve the performance of SDR systems.