Research On Synchronization Techniques In High Dynamic Communication Links Of LEO Satellites

The research on synchronization techniques in high dynamic communication links of LEO satellites is taken in this paper. The difficulties of modern satellite communication synchronization technology and its development trend are analyzed. And on this basis, the synchronization problem in LEO satellite high dynamic communication link is refined into three sub-problems: namely, carrier frequency acquisition with low SNR, large frequency offset, and fast time-varying frequency; carrier phase tracking with low SNR and fast time-varying frequency; fast burst frame synchronization under low SNR circumstance. The main research contents and innovative results are as follows:(1) Frequency modulation signal’s spectrum flattening effect is analyzed, and the concept of "optimal detection interval" is proposed. Besides, the theoretical expression of the optimal detection interval and the detection performance is derived. The problem of insufficient resolution of standard discrete Chirp Fourier transform(DCFT) algorithm used for high dynamic communication link carrier frequency acquisition is discussed, and accordingly the improved MDCFT algorithm is proposed, which introduces a fraction factor into the Chirp basis function and refines the resolution of the target area. For linear frequency modulation(LFM) signal, the acquisition algorithm based on MDCFT is proposed, and the computational complexity is analyzed. On this basis, the VS-MDCFT algorithm with higher computational efficiency is given, which could sharply reduce the computing complexity about 60%. For NLFM signal parameter estimation, the applicability of the optimal detection interval is shown, and the polynomial Chirplet transform with optimal detection window(ODW-PCT) is used for the acquisition. Theoretical analysis and simulation show that, the performance of ODW-PCT algorithm is superior to general PCT algorithm.(2) The signal model of the carrier phase tracking in LEO satellite high dynamic communication link is established and the general structure of carrier phase tracking based on Bayesian estimation theory is proposed. Limitations of the standard Kalman filter applied to carrier phase tracking are analyzed. For the selecting of the importance sampling density function of particle filters, the likelihood sorting sampling algorithm based on strong incentives and augmented set of particles is proposed, which could get more reliable particles close to the posterior probability distribution by increasing incentives noise power and expanding the scale of the particle set. At the same time, the latest measurement data is introduced into the particle generation process through the likelihood value ranking mechanism. Compared with the Kalman filter algorithm and standard particle filtering algorithm, the new method could obviously improve the tracking accuracy and stability.(3) Given the impact of the system state space model on carrier tracking system based on Bayesian estimation, demonstrating the limitation of high-order truncated model used in the general carrier tracking system. The feasibility and advantage of the maneuvering target modeling technique applied to high dynamic tracking system is discussed, and the 4-D current statistic(CS) state space model used in carrier tracking is proposed. Under the Jerk environment, the tracking performance is improved about 1 d B compared with the general carrier tracking algorithm using the higher order truncated model. Illustrating that carrier tracking algorithms with single dynamic model could not achieve optimal performance when used in the high dynamic communication link with fast time varying relative motions. Giving the carrier tracking algorithm based on multimode fusion, the basic idea is to establish multiple parallel work estimators for the carrier tracking module and configure different estimators for different maneuvering target models. In this way, the final carrier phase tracking value is combined by all of the models. If pre-allocated estimators is much enough and the selected model range is wide enough, the carrier tracking system can always guarantee the accuracy and stability in the whole communication time.(4) The basic structure of the frame synchronization based on sliding window detection is proposed. For uncoded frame synchronization, the probability of false alarm and miss alarm are deduced based on hard decision and soft decision respectively, showing the poor detection performance of uncoded frame synchronization algorithm under low signal-to-noise ratio circumstance. For coded frame synchronization, a coded auxiliary burst frame synchronization algorithm based on detecting check matrix of LDPC code is proposed, which could work in low SNR circumstance. And the detection performance is greater than uncoded frame synchronization algorithm. The main problem of the algorithm is the pretty high computational complexity. To solve this problem, this paper gives an effective LDPC coded burst frame synchronization algorithm based on the non-convex optimization threshold searching. The basic idea is dividing the detection progress into two steps, and detecting the boundary through SW correlation and LDPC code word check relationship respectively. The second step detection uses variable operation mode, and its computational complexity is directly related to the measurement of first step, which could ensure that the final detection performance to meet requirements and compress redundant computation as much as possible so as to improve the efficiency of algorithm. When designing the algorithm progress, the key threshold setting is modeled as a mathematical non-convex optimization problem. Through finding several local optimal solutions within the feasible region, the algorithm could(approximately) achieve the minimum computational complexity while ensuring that the algorithm performance, such as probability of false alarm, miss alarm, could meet the expected requirement in theory. Simulation results show that the efficient LDPC coded burst frame synchronization algorithm based on the non-convex optimization threshold searching can work in low SNR with good detection performance(Es/N0 =- 3d B, the Pf = Pm= 4x10-8). Compared to other algorithms, this algorithm has obvious advantages in detection performance aspects, such as the probability of false alarm, miss alarm, and the computational complexity.