Research On System Of Hyperspectral Microwave Radiometer Based On FPGA For Detecting Atmospheric Temperature

Microwave radiometer is one important tool to detect very important atmospheric parameters which are temperature and humidity.For vertical detection of atmospheric temperature and atmospheric humidity,the number of detection channels is related to spectral resolution and vertical resolution.In recent years,the number of frequency detection channels of microwave radiometers has been increased by using hyperspectral technology to achieve higher spectral resolution and vertical resolution.This technology also has an important application value in the area of microwave atmospheric measurement,especially in the area of narrow-band trace gas sounding.The back end of this kind of microwave radiometer can be implemented with different technical forms,such as filter banks spectrometer,digital spectrometers(including Fast Fourier Transform spectrometer and auto-correlation spectrometer),chirp transform spectrometer and acousto-optical spectrometer.In this paper,by comparing and analyzing the development of hyperspectral microwave radiometers,a system design of hyperspectral microwave radiometer is proposed and the corresponding research popurse is to improve spectral resolution of microwave radiometer and vertical resolution of atmospheric detection.It can reduce complexity of the system,improve universality of the back end of the receiver.This thesis presents the whole work of design and completion of this hyperspectral microwave radiometer,focuses on the reaserch of key technologies,as well as concentrates on hardware algorithm in FPGA to gain spectrum from parallel multichannel,and shows the atmospheric measurement test.The main contents include:Firstly,by analyzing and researching the structure of microwave ratiometer,a ground-based wideband hyperspectral microwave radiometer system design based on Field Programmable Gate Array(FPGA)and spectral analysis technology is proposed.This system has one broad band channel with two mixers.It can measure atmospheric information over 50-60 GHz and its output signal has frequency bandwidth of 2GHz by changing frequencies of local oscillators five times,which decreases the complexity of the whole hyperspectral microwave radiometer.Fast Fourier Transform(FFT)spectrometer is the back end of this system.It is the key module of this hyperspectral microwave radiometer to achieve spectral analysis function by using wideband digital sampling and FPGA technologies.However,traditional microwave radiometer is a super-heterodyne architecture and has a backend with comb filter banks to gain atmospheric information through around 5-13 sounding channels.Traditional technologies are not good enough to acheive high spectral resolution and vertical resolution due to few frequency channels and low-frequency digital sampling technology.There are three advantages.Firstly,it is one signle-channel wideband receiver which reduces the complexity of the system.Secondly,it is a spectrum analysis receiver with a high speed digital FFT spectrometer whose spectral resolution of 10 MHz over 50-60 GHz which can improve the vertical resolution.The last one is that its number of sounding channels can be changed.Secondly,wideband microwave radiometer receiver has been completed by using broadband and high speed sampling technology and spectrum analysis technology in FPGA.FFT spectromer realizes the spectrm analysis function through the cooperative work between data acquisition board and the controlling and transforming board.It has characteristics of broad frequency band input signal and lots of complex signal lines.In hardware implementation,an adjustable analog delay chip is used to ensure that the delay between synchronization signals is adjustable.Multi-layer is used to reduce the difficulty of routing and crosstalk among different lines.At the same time,keeping differential routings having strictly equal length is used to ensure the synchronization of differential signals.Thirdly,this paper propses a multi-channel high-speed digital spectrum analysis algorithm based on FPGA to solve the difficulty of processing parallel multi-input highspeed data.This method changes the sigle-channel input-output structure of the traditional polyphase filter structure to parallel multi-channel input-output.This new structure meets the requirements of parallel multichannel inputs in the backend of this hyperspectral microwave radiometer.At the same time,this algorithm can adjust its own structure to achieve adjustment of the output data rate to adapt to the different processing rates inside FPGA.Improved polyphase filterbanks structure can deal with parallel 64-input and output 8 parallel signals at speed of 156.25 MHz.A complex parallel Fast Fourier Transform structure after polyphase filterbanks structure can achieve spectrum of 2N real points of parallel signals.It is achieved by computing N complex points Fast Fourier Transform,which saved a lot of resources in FPGA.This spectrometer works continuously and has more than one thousand channels over frequency bandwidth of 10 GHz.At last,we set up this hyperspectral microwave radiometer system to carry out atmospheric brightness temperature experiment over 50-60 GHz in January,2020.And the bright temperature distribution curve of the atmospheric oxygen absorption spectrum show an upward at 50-55 GHz and flatten at 55-60 GHz,which is basically the same as old measurement results of clear sky in winter.At the same time,the algorithm proposed in this thesis is verified to be used to achieve atmospheric bright temperature,and the error analysis of experimental results is given.