Optimization And Implementation Of 5G NR Physical Layer Channel Estimation And MIMO Detection Based On GPP

The physical layer is the core of all communication technologies.Its vigorous development in the past few decades is the key to promoting the evolution of mobile communications from the first generation(1G)to the fourth generation(4G).Today,facing emerging scenarios marked by the interconnection of all things,driverless and AR/VR technologies,mobile communication is moving towards the fifth-generation mobile communication technology(5G)with high bandwidth,large connection,low latency and high reliability.Evolution.In order to achieve the ultra-low latency requirements of 5G communication technology,it is the current research trend to optimize the real-time nature of baseband signal processing on General Purpose Processors.This paper mainly studies the optimization and implementation of 5G NR physical layer channel estimation and MIMO detection based on GPP.After fully investigating the channel estimation and MIMO detection algorithms,the link-level simulation of the physical downlink shared channel of 5G NR was first implemented on the Matlab platform to verify the feasibility and correctness of the algorithm.Then combined with the execution characteristics of the GPP platform to design an algorithm parallelization scheme,disassemble and subdivide the complex formula of the algorithm,optimize the algorithm execution structure and sequence,reduce the dependence between data,and use the AVX-512 instruction set to achieve fine-grain The data is parallel,which speeds up the calculation speed,reduces the calculation time,and improves the system throughput.Finally,according to the simulation results,when the antenna is configured for 2 transmit and 2 receive,the delay is reduced by nearly 20 times after using the AVX-512 instruction set,which can meet the needs of the project.