Research On The Techniques Of Satellite Signal Detection And Recognition

As an important part of modern communication satellite communications own a veryimportant position in the civilian and military fields. In recent years, with the continuousdevelopment of communication technology, there are some new difficulties in satellite signalinterception and analysis. Due to the influence of receiving technology, geographic location andpoor channel environment, regional satellite signal detection faces great difficulties innon-cooperative receiving environment. With the increasing of the non-standard and unknowninstitutional signals， more and more new modulation schemes were used in satellitecommunications. Among them, APSK signal has been widely used because of its higherspectrum utilization compared to PSK and better nonlinear adaptability than QAM. But researchof APSK modulation classification is less. This paper researches on solving the two problemsabove: firstly, to improve the existing algorithm to adapt to searching needs of a specific channelenvironment; secondly, to analyze the APSK signal’s characterization after completing signalsearch, and to identify the commonly used phase-modulated signal (PSK, QAM, APSK etc). Themain contribution can be summarized as follows:1. For the ineffective of traditional energy threshold search method to signal detection undera specific channel environment, whose signal spectrum fluctuates largely, different fromtraditional satellite signal. On the one hand, researches optimal spectral estimation method, themultitaper spectral estimation method with approximate unbiased and low variance is introducedto solve the problem of bias and variance dilemmatic for spectral estimation; on the other hand,adding windows based on the large-scale wavelet transform suppressing noise to suppress noiseagain, can improve the traditional wavelet transform singular point detection performance.Combine the advantages of both above, and select an appropriate search threshold to enhance thesearch performance of the satellite signal in a specific channel environment.2. The existing modulation recognition algorithms of APSK mostly need matched receivingand strict preprocessing. In the environment of non-cooperative reception, it is hard to reach thepretreatment requirements of the algorithms and leads to a serious deterioration of therecognition for short of prior knowledge. This paper proposes a recognition algorithm based oninstantaneous statistics and signal-to-noise ratio estimation, which is based on the band receivedsignal, low complexity, less dependent on prior knowledge and pretreatment requirements. Then,the instantaneous statistical characteristic is demonstrated not sensitive to the noise.3. This paper theoretically analyzes and deduces the high-order cyclic cumulants expressionof PSK, QAM, APSK signals after a non-linear transformation. A recognition algorithm isproposed based on the different high-order cyclic cumulants discrete spectral characteristics and cyclostationary detection to distinguish QPSK,8PSK,16QAM,16APSK and32APSK. Thismethod does not require preprocessing task such as estimation of parameters andsynchronization, and can effectively resist the impact of noise and fading channels. Thesimulation results verify the correctness of theoretical analysis and efficiency of classificationalgorithm.4. A specific satellite channel signal search and APSK signal recognition analysis system isimplemented based on the relevant theoretical research, and the relevant theories as well asalgorithms are also tested on a general signal processing platform. Engineering implementationand experimental testing results validate that the system is pragmatic.