Research And Implementation Of LEO Satellite Signals Combining Techniques Based On Multi-Antenna

LEO satellite communication systems have shown great prospect due to the low orbit, shortcommunication delay, low link loss and flexible networking, LEO satellite movement speed, so asingle large aperture antenna tracking receiver difficulties, utilizing multiple small-diameterantenna arraying instead of large-diameter antenna, the satellite signal tracking receiver, whilealso reducing the costs. Because of the variety of affects during the LEO satellite signaltransmission, the receiver’s Signal-to-Noise Ratio (SNR) is low, while utilizing multi-antennasignal combining techniques, by the number of antenna acceptance the signal from the samesatellite, then eliminate the difference of parameters between the signals, and the weightedcombine all signals based on the characteristics about the sameness of the signals’ componentand the non-correlation for noises’ component. can effectively improve the quality of thereceived signal.Because there is relative motion between the LEO satellite ground receiving antenna,resulting in the time-delay difference and the carrier frequency difference between the signals onthe non-linear changing, and further caused by the phase difference time-varying, creating newproblems to estimating the parameter differences of signal combining techniques. On the basis ofin-depth study of randomly-distributed multi-antenna signal combining techniques, for how toreduce the computational load of the joint time and frequency offset estimation and time-varyingphase difference estimation and compensation have done depth discussion and analysis, designedand implemented a multi-antenna signal combining software. The contributions obtained in thisthesis can be summarized as follows.1. Aiming at the time delay difference and frequency difference estimation of themulti-antenna signal combining techniques, at first, analyzed the time-varying time delaydifference and frequency difference; Followed makes a further research and analysis on how toreduce the computational load of the joint time and frequency offset estimation based on thetwo-order cross ambiguity function, transfer the cross ambiguity function peak search to thenonlinear unconstrained optimization, utilizing an improved simplex algorithm to achieve theglobal optimum search, simulation results show that, the improved simplex algorithm can doesnot affect the estimated accuracy, which greatly reduces the computational load of the joint timeand frequency offset estimation.2. Aiming at estimate and compensate the time-varying phase difference based on twosignals, at first, utilizing Cross-correlation algorithm estimate of the time-varying phasedifference, the result show that the theoretical estimated value equal to the center sampling pointphase difference of the integral data; And analysis the reasons of phase difference poor estimateaccuracy caused by residual frequency difference; Followed analysis of the integral SNR gainand the theoretical estimation performance, given the method of calculation of the optimalintegration length of the data, when the integral length of the data and the normalizedresiduafrequency difference of the product is about2.33, have the highest phase estimationaccuracy; Finally, comparative analysis the synthetic performance of two different package about estimate and compensate the time-varying phase.3. Aiming at estimate and compensate the time-varying phase difference based on multiplesignals, at first, have the theoretical derivation of SUMPLE algorithm based on the time-varyingphase difference, and analysis of the effects on synthesized signal to noise ratio and the estimatedperformance of the combining weight caused by the residual frequency difference; TheSUMPLE algorithm has better synthetic performance than the SIMPLE algorithm, and for moresignals for better. Followed proved the convergence characteristics of the SUMPLE algorithmbased on time-varying phase, and analysis of the phase convergence value of the synthetic signal;The last, analysis of the synthetic performance of the estimate and compensate package of pointby point sliding, and simulation data shows that this method could guarantee the sustainedconvergence of the SUMPLE algorithm, and could make the synthesized signal to noise ratio tominimize the damage caused by residual frequency.4. According to the project requirements, design and implement a multi-antenna signalsynthesis software. The structure of the software is divided into five modules, which are theinterface control and display module, the data reading and writing module, the joint time andfrequency offset estimation module, the data processing module and the signal combiningmodule. Using the VC++and Matlab mixed programming mode complete the preparation of thesoftware program. The results of simulation and testing indicate that the performance of thesoftware is good and has good versatility and scalability.