Research And Implementation Of 5G Terminal PDSCH Channel Optimization Based On OAI

In recent years,the extensive deployment of The 5th Generation Mobile Communication System(5G)network is bound to bring huge economic pressure on equipment,and the incompatibility between various equipment manufacturers is becoming more prominent.Therefore,it is urgent to seek an open and sustainable network construction scheme.With the improvement of computer processor performance in recent years,it has become a new idea to construct 5G network using Software Defined Radio(SDR)technology.Open Air Interafce(OAI),as a famous SDR 5G open source platform in the industry,has been used by various institutions and universities as the basic platform for algorithm research and secondary development.Physical Downlink Shared Channel(PDSCH),as a service data bearing channel in 5G network,occupies the majority of data delay.How to improve the efficiency of PDSCH channel has become the key to improve the performance of 5G network system.Based on OAI platform,this paper analyzes the PDSCH channel of OAI platform,and proposes optimization schemes for the deficiencies of terminal PDSCH channel.Specific work is carried out as follows:(1)Based on the general computing platform of X86 architecture,this paper builds the5 G Radio Access Network system(RAN)of OAI platform,analyzes the PDSCH channel architecture of OAI platform,and gives the relevant performance of the PDSCH channel of the base station and the terminal under different transport block size.The results show that the relevant performance of the base station PDSCH channel is well,while the terminal PDSCH channel maintains well Cyclic Redundancy Check(CRC)error rate,but the maximum slot time is 409.4% of the theoretical value.The maximum difference between the downlink peak rate and the theoretical value is 66.1%.Based on the above problems,this paper further analyzes the terminal PDSCH channel,and the results show that the total time consuming of the unscrambling module and channel decoding module occupy 89.3% of the terminal PDSCH channel and 74.3% of the terminal time slot.After analysis,it was found that the generation of the unscrambling sequence and the unscrambling of the transmission block in the unscrambling module are highly coupled.In the channel decoding module,serial decoding is used for multiple code blocks,resulting in a problem of high data latency in the OAI platform terminal PDSCH.(2)In view of the time-consuming problem of PDSCH channel unscrambling module and channel decoding module of OAI platform terminal,this paper proposes an optimization scheme of PDSCH channel based on thread pool.The scheme uses the thread pool to complete the decoupling and parallelization design of unscrambling sequence generation and transport block unscrambling,and completes the parallel decoding processing of multiple code blocks in the channel decoding process stage.Although CRC error rate and system power consumption has been increased,but the time consumption of the unscrambling module and the channel decoding module is effectively reduced,and the peak downlink rate is increased.The experimental results show that,under the premise of meeting the CRC error rate of the protocol specification,compared with the original version of the OAI platform,the time consumption of the unscrambling module is reduced by 86.3%,the time consumption of the channel decoding module is reduced by 86.1%,the time consumption of PDSCH channel is reduced by 76.5%,and the time consumption of terminal slot is reduced by 63.3% and the downlink peak rate is increased by 131.7% at most,which effectively improves the performance of the OAI platform.(3)In view of the time-consuming problem of PDSCH channel unscrambling module and channel decoding module of OAI platform terminal,this paper proposes a terminal PDSCH channel optimization scheme based on multiple Graphics Processing Unit(GPU).The scheme first reuses the processing of thread pool for unscrambling sequence generation,and uses GPU Compute Unified Device Architecture(CUDA)to complete the data division and parallel processing design of unscrambling module and channel decoding module at bit level and code block level respectively under multiple GPUs.Although CRC error rate and system power consumption of the system have also been increased,but the time consumption of the unscrambling module and the channel decoding module is effectively reduced,and the peak downlink rate is increased.The experimental results show that,under the premise of meeting the CRC error rate of the protocol specification,compared with the original version of the OAI platform,the time consumption of the unscrambling module is reduced by 92.1%,the time consumption of the channel decoding module is reduced by 83.7%,and the time consumption of the terminal PDSCH channel is reduced by 77.8%,the terminal slot time consumption is reduced by 64.2%,and the downlink peak rate is increased by 138.3%at most,which effectively improves the performance of the OAI platform,and most of the performance of this scheme is better than that of the previous scheme.