Research And Implementation Of Software Defined Radio Digital Spectrum Analysis Technology Based On SOPC

The application of 5G will bring opportunities to the development of IoT,and the massive access of the IoT will aggravate the shortage of spectrum resources.Cognitive radio scans the spectrum status of accessible bands through radio spectrum analysis,and dynamically accesses idle frequencies to improve spectrum utilization efficiency.Radio spectrum analysis technology is a key technology for cognitive radio and a prerequisite for cognitive radios to discover idle frequencies.At the same time,radio spectrum analysis is also very important in radio monitoring.Its purpose is to ensure the normal operation of legitimate wireless communication systems and to search for illegal communications and locate illegal communication equipment.It has a high status in every country.Both the IoT and radio spectrum monitoring require the portability and low power consumption of spectrum analysis equipment.The use of SOPC can make the equipment miniature and low power consumption.In the thesis,the radio spectrum analysis technology,including digital signal processing techniques such as sample rate conversion,Fourier transform and spectrum detection,is researched.Based on the research of spectrum analysis technology,a radio spectrum analysis system is designed and implemented on SOPC.The radio spectrum analysis system is implemented on SOPC.On the SOPC,CPU provides flexible configuration for the system and FPGA executes high-speed calculation in configurable spectrum analysis module.Due to system throughput and measurement requirements,following improvements have been made to the relevant modules in the spectrum analysis system:1)Improving the FFT ping-pong RAM implementation structure to improve FFT throughput;2)Improving block floating point FFT implementation,each time the data of the middle butterfly operation determines the decimal point position separately,and improves the small signal power precision in the multi-frequency signal;The simulation shows that when there are two signals and their power difference exceeds 30 dB,the power accuracy of small signal with the improved block floating-point FFT measured is increased by an order of magnitude,while the power accuracy of the large signal is not affected.3)A segmentation fixed-point strategy is adopted for the window function coefficients to ensure higher precision of window function processing.According to the system design,spectrum analysis module is implemented and vertified in the thesis.40% throughput is increased but only 28.6% BRAM is added in the improved ping-pong RAM structure.The improved block floating point FFT data processing mode and the window function is truncated by segmentations do not make the BRAM and DSP increase.