Key Technologies Of Spaceborne Communication Signal Processing For LEO Satellite

Driven by the development of satellite technology, the trend towards network andformation is increasingly deepened for LEO satellite. The LEO satellite network thatcomposed of inter-satellite links can provide support for the global data transmission and avariety of applications. Therefore, LEO satellite network has received many attentions frommilitary and scientific research departments around world. Transformations on applicationand role propose a series of challenges for LEO satellite communication system. From theperspective of communication signal processing, these challenges mainly include enhancethe precision of acquisition, reduce the cost of signal processing, improve the powerefficiency and so on. Based on the LEO satellite system having inter-satellite links andon-board processing capability, to address the challenges faced by LEO satellite spacebornecommunication signal processing, around the above aspects this paper makes research onthe following techniques: DSSS signal high precision acquisition, sparse simplifiedtime-frequency signal processing, and the most compact high-order modulation. Thispaper’s main work and innovative achievements are as follows.This paper presents the high-precision frequency rearrangement acquisition method, inwhich the two-dimension estimation is simplified into one-dimension problem byintroducing phase-frequency characteristics. Different form traditional methods, thisalgorithm makes full use of signal’s phase-frequency characteristics. As the two parametersare presented by phase-frequency and amplitude-frequency characteristics respectively, byconsidering the constraint between them, the two-dimension estimation is simplified intoone-dimension problem so as to be solved through one operation. The use ofphase-frequency characteristics improves the accuracy of this algorithm without increasingacquisition time. In this paper, influencing factors are studied, anti-noise performance isanalyzed, non-compact SNR theoretical threshold is deduced, and acquisition accuracy issimulated for this algorithm. Results show that the time and frequency accuracy of thisalgorithm are improved by50%and60%compared with traditional method.This paper presents the frequency rearrangement combined decoupling acquisitionmethod, in which the effect of amplitude-frequency characteristics on phase-frequencycharacteristics is immobilized so as to reduce limitations on either process. The coupling oftime and frequency estimation processes has influenced anti-noise performance in the high-precision frequency rearrangement acquisition method. By introducing combineddecoupling method into rearrangement process, the features of amplitude-frequency andphase–frequency characteristics are maintained meanwhile their estimation processes areuncorrelated. Such decoupling treatment reduces the limitations on each process andimproves the overall anti-noise performance without increasing acquisition time. In thispaper, anti-noise performance is analyzed and non-compact SNR theoretical threshold isdeduced for this combined decoupling algorithm. Results show that the threshold isimproved by about6dB compared with the high-precision frequency rearrangementacquisition method.This paper presents the positioning optimized sparse Fourier transform algorithm, inwhich the computational complexity of time-frequency sparse processing is reduced bymaking full use of the band-limited sparse feature of DSSS signal. For the traditional sparseFourier transform method, sparse processing is essentially the solving of underdeterminedequations, thus the steps of compressing, solving and selecting need to be carried outsequentially. Different form traditional method, this algorithm prevents the collision ofeffective spectrum peaks by making full use of the band-limited sparse feature of DSSSsignal. In this algorithm, sparse processing can be transformed into the solving ofconventional equations with fluctuation results, which makes it possible to arrange theselecting step before solving step. Such arrangement improves the overall computationalcomplexity without reducing the accuracy of estimation results. In this paper, thepositioning optimized sparse Fourier transform algorithm is proposed and analyzed, thenapplied into the former described acquisition methods. Results show that the computationalcomplexity of this algorithm is reduced by about50%compared with the original sparseFourier transform algorithm, and by applying this algorithm the computational complexityof former described acquisition methods can be reduced by about96%and90%respectively.This paper establishes a generic mathematical model for the most compact high-ordermodulation method, based on which the generic analytical expression of anti-noiseperformance is deduced and a low complexity demodulation algorithm is proposed.Restricted by the imperfection of former models, previous studies mainly focused on twoaspects: theoretical study on the performance of constellation on infinite plane andexperimental analysis of constellation with certain points. To solve the above problems, thispaper presents a constellation points magnitude classification based generic model for the most compact high-order modulation method. Based on the proposed model, the genericexpression of anti-noise performance is deduced and the three-phase demodulationalgorithm with constant complexity is proposed. Results show that the model andexpressions are consistent with actual situation, this method’s modulation efficiency ishigher than that of QAM in large constellation, and the proposed demodulation algorithmhas a constellation-order-independent complexity of18real multiplications plus9realadditions.