Research On GPS Date Denoising Based On EMD

As the improvement of technology and times, GPS is more and more important in our daily life with it’s advantages such as all-weather, high precision, high efficiency and use easily. Navigation of ships, cars and planes, urban intelligent traffic management, emergency rescue, outdoor tourism and explore all couldn’t work without GPS. All these productions we use are derived from the solution of the GPS signals. So the quality of results we got from GPS is determined by the accuracy of GPS signals. GPS signals can be affected by many kinds of error sources, such as ephemeris error, satellite clock error, ionospheric delay, tropospheric delay, receiver clock error and so on, which cause deviations between GPS signals and real signals. These errors also have certain influence on results’ accuracy in signal solution processing that could made inconvenience to daily use and research. So the GPS signals denoising is very necessary.The main works completed as follows:(1) Firstly is the analysis of the existing time-frequency analysis method. There is not only made a summarize for the classification and conception of traditional time-frequency method such as Fourier transform, wavelet transform and Winger-Ville distribution, but also talked about their advantages and disadvantages. Then introduced the EMD from instantaneous frequency, intrinsic mode function, algorithm flow, Hilbert spectrum, marginal spectrum these five aspects and illustrate them with a simulated example.(2) Secondly is the discussion about EMD and proposed a new method to solve the end effect. This part analyzed the empirical mode decomposition in detail, which is proposed by Chinese American Norden E.Huang, and explained it’s basic principle and decomposition process. After the signal decomposition based on EMD, we got a set of IMF(Intrinsic Mode Function) and a trend component. At last a new method to deal with the end effect is proposed based on the discussion of EMD’s characteristics and problems. These problems include lack of theoretical basis, end effect, mode mixing and so on.(3) Third is denoising analysis of GPS signal. For the problem that solution precision can be affected by noises in GPS signals, we put CEEMD(Complementary Ensemble Empirical Mode Decomposition)in use of the denoising of simulated GPS signal and GPS double difference observation, and made a denoising results’ comparison between EMD, EEMD(Ensemble Empirical Mode Decomposition) and CEEMD with three standards: SNR(Signal-Noise-Ratio), mutual correlation coefficient( R) and mean square error(MSE). The comparison results show that CEEMD is much better than EMD and EEMD.