| The navigation signal that is reflected by physical surface can carry abundant surface information.GNSS-R mainly studies the reflected signals of navigation satellites,it can set up a corresponding mathematical model to deduce physical parameters and all kinds of information on the surface.Compared with the traditional radar remote sensing detection,GNSS-R has a advantage of all-weather and high space-time resolution.It has a wide range of applications in wind speed and wind direction,effective wave height,height measurement,soil moisture,salinity of sea water and typhoon warning.The traditional reflected signal software receiver captures and tracks the direct signal,and then demodulates the navigation message to get Navigation informations such as current receiver position,where the processing of the reflected signal needs processing of multiple code delay correlation,and inverses the results of the retrieved direct and inverse processing,to obtain the remote sensing information.Because of the large amount of parallel computation,the processing time of the software receiver is greatly limited compared with the traditional hardware receiver processing.Parallel processing based on GPU can not only greatly improve the processing efficiency,but also can adapt to a variety of new modulation navigation signals relative to the traditional hardware architecture.This paper studies many inverse models in the wind detection,the effect of interference effects on ocean altimetry accuracy and the new concept of measurement of navigation signal based on the detailed signal reflection principle,characteristics,direct signal and reflected signal processing methods.The paper is designed by GPU+CPU architecture based on heterogeneous reflected signal processing,which uses satellite integrated streaming capture of direct signal,to get a rough code phase and carrier value,and processes the direct signal by way of direct parallel protocol.In the open-loop tracking mode,a hierarchical coarse granularity protocol is designed to calculate the related power of the reflected signal.At the same time,the host computer software is developed to receive the processing result from the hardware architecture,including the serial port to receive the navigation chip UM220 data,the USB port to accept the original data collected by the acquisition card and the correlation value of the reflected signal.Finally,the parallel processing software of sea surface electromagnetic scattering signal model is developed.where the QT is used as the interface,and the underlying GPU processing method is used to calculate the theoretical correlation power values.The simulated test data and the original data are processed.The experimental results show that the parallel processing mode based on GPU can greatly improve the processing efficiency,and lay the foundation for real-time processing and inversion of reflected signals. |
