Research On Wireless Channels Adaptive Equalizer Based On SOC

With the development of information technology,the increase in communication protocols and the increase in bandwidth requirements,software radio has received more and more attention.Higher requirements are put forward for the multi-function,integration,scalability and high speed of the software radio platform.When broadcast signals encounter areas such as high-rise buildings,underground tunnels,etc.,the signal will be attenuated and even blind spots may exist.Therefore,it is necessary to perform power equalization on broadcast signals in various frequency bands before amplification,which is conducive to signal optimization and transmission.Aiming at the problems that traditional radio platforms cannot handle high-speed signals,the signal quality is reduced due to unbalanced power during signal transmission,and is not conducive to subsequent processing,a SOC-based wireless channel adaptive equalizer is designed.The main research work is as follows:(1)Aiming at the shortcomings of traditional radio systems,a software radio platform based on AD9361 radio frequency agile transceiver and ZYNQ processor is proposed.The platform adopts a hybrid architecture of FPGA+ARM,connecting various modules and custom logic functions as required.The ARM end of the ZYNQ processor controls all peripherals and memory controllers to ensure the independent operation of the subsystem.There are a large number of expandable modules on the FPGA side,which can fully define the I/O interface,provide an internal bandwidth of more than 100Gb/s,and integrate a high-speed serial interface.The FPGA and the ARM share part of the memory,which can complete high-speed data interaction.(2)In order to reduce the occupation of internal resources of the ZYNQ processor,an ICAP(The Internal Configuration Access Port)download method is designed.Compared with the traditional platform,there is no need to replace the hardware,just set the starting address and boot program inside the system,and realize the online update of the system through the developed QT download interface.(3)Aiming at the problem of power imbalance in the process of broadcasting signal transmission,two equalization algorithms of BP neural network and improved LMS are analyzed and compared.With the help of MATLAB simulation platform,the impact of the two algorithms on the system bit error rate is analyzed under the same conditions.According to the simulation results,the improved LMS equalization algorithm is better than the neural network algorithm in terms of system convergence speed and bit error rate,so the improved LMS algorithm is selected to map to the hardware equalization algorithm.A comparative analysis of the error generated by the equalization algorithm in FPGA and MTATLAB,the errors of the two are basically the same,and the algorithm can be transplanted.Due to the limitation of hardware resources,the structure of the equalizer cannot be infinitely long.Through MATLAB simulation,it is determined that the optimal structure of the equalizer is 7 steps,and the step value is 1/128.(4)The equalization algorithm is designed into an IP core module with the help of the VIVADO platform to complete the design and development of the equalization algorithm.The functional simulation was completed by MODELSIM,which verified the feasibility of the design.Analyzing the consumption of resources,the register occupies 8% of the total resources and the memory occupies 10%,which greatly saves resources.The equalization module is integrated into the system,and the time domain and frequency domain waveforms before and after equalization are observed through ADI's IIO-Oscilloscope software.Through comparative analysis,no matter the original signal is high or low,after adaptive equalization,the average power always remains within the range of 50 d BFS to 56 d BFS,realizing multi-band adaptive power equalization.This article has certain reference significance for the design of SOC-based software radio platform.