低轨卫星高动态通信链路同步技术研究

发布时间：2018-04-25 10:57

本文选题：低轨卫星 + 高动态　；参考：《北京理工大学》2015年博士论文

【摘要】：本文针对低轨卫星高动态通信链路中的同步技术展开研究,分析了现代卫星通信同步技术领域面临的主要难点及其发展趋势,并以此为基础,将低轨卫星高动态通信链路中的同步问题细化提炼为三个子问题:即低信噪比、大频偏、快速时变频率条件下的载波频率捕获问题;低信噪比、快速时变频率条件下的载波相位跟踪问题;低信噪比条件下的快速突发帧同步问题。本文的主要研究内容和创新性成果如下:(1)分析了线性调频(LFM)信号的频谱扁平化效应,并提出LFM信号“最佳检测区间”的概念,推导了最佳检测区间的理论表达式及其检测性能;论述了标准离散Chirp傅里叶变换(DCFT)算法用于高动态通信链路载波频率捕获时存在分辨率不足的问题,并据此提出改进的MDCFT算法,通过引入分数因子至Chirp基函数中,细化目标区域的分辨率;针对LFM信号参数估计问题,提出基于MDCFT的频率-调频率二维空间最大似然搜索算法,分析了算法的计算复杂度;在此基础上,提出计算效率更高的基于可变规模MDCFT(VS-MDCFT)的参数估计算法,将原本计算量消耗巨大的MDCFT拆分为两个规模较小的子MDCFT,在保证捕获概率没有明显降低的前提下缩减了约60%的计算量。针对非线性调频(NLFM)信号参数估计问题,论述了“最佳检测区间”对NLFM信号的适用性,并据此提出基于最佳检测窗多项式Chirplet变换(ODW-PCT)的参数估计算法。理论分析和仿真表明,配有最佳窗函数的ODW-PCT算法的捕获概率明显优于一般PCT算法。(2)推导了低轨卫星高动态通信链路在载波相位跟踪阶段的信号模型,给出基于贝叶斯估计理论的一般载波相位跟踪结构。分析了标准卡尔曼滤波(KF)应用于载波相位跟踪系统的局限性,并推导扩展卡尔曼滤波(EKF)和粒子滤波(PF)用于载波相位跟踪时的一般信号处理流程。从粒子滤波重要性采样密度函数的选取问题出发,提出强激励增广粒子集似然排序采样算法,通过增强激励噪声和扩大粒子集规模两方面措施,获得更多数量的贴近于后验概率分布的可靠粒子;同时,通过似然值排序机制将最新测量数据信息引入粒子生成过程中。理论分析和仿真结果表明,基于强激励增广粒子集似然排序采样的粒子滤波算法,其跟踪稳定性和准确性较卡尔曼滤波类算法和标准粒子滤波算法有明显提高。(3)分析了系统状态空间模型对基于贝叶斯估计的载波跟踪系统的性能影响,论证了一般载波跟踪系统所用高阶项截断模型的局限性。借鉴雷达目标跟踪领域中的机动目标建模技术,从多普勒效应的产生机理出发,论证了机动目标建模技术应用于高动态载波跟踪系统的可行性和优越性,并提出基于四维CS状态空间模型辅助的载波跟踪算法,其在Jerk环境下的跟踪性能比采用高阶项截断模型的一般载波跟踪算法改善约1d B。论述了采用单一动力学模型的载波跟踪算法无法在具有快速时变相对运动状态的高动态通信链路中取得最优性能,并以此为背景,提出基于多模融合的载波跟踪算法,其基本思想是为载波跟踪模块建立多个并行工作的子估计器,且不同估计器配置不同的机动目标模型,跟踪器最终输出的载波相位跟踪值由各子估计器的输出值按一定准则联合计算获得。若预分配的子估计器数量足够多,所选模型覆盖范围足够广,则对于通信载体在任意时刻的实际运动状态,均有很大可能找到与其相匹配的动力学模型,使得载波跟踪系统在整个通信持续时段内始终能够保证载波相位跟踪的精确度和稳定性。(4)提出基于滑动窗检测的突发帧同步系统基本结构。对于无编码辅助帧同步,分别推导了基于硬判决和软判决帧同步字段(SW)检测的帧同步算法的虚警、漏警概率,论述了无编码辅助帧同步算法在低信噪比条件下检测性能较差的问题。对于编码辅助帧同步,提出一种基于校验矩阵检测的LDPC码辅助突发帧同步算法,能够工作于低信噪比条件下,其检测性能远高于无编码辅助帧同步算法,但存在计算复杂度过高、不易于实时处理的缺点。为了解决这一问题,提出一种基于非凸优化门限搜索的LDPC码辅助高效突发帧同步算法,其基本思想是将帧同步检测分为两级,分别利用SW相关性和LDPC码字校验关系联合检测帧边界。其中第二级检测采用可变运算量模式,且其运算量大小与第一级检测所得度量值直接相关,从而在保证两级处理算法的检测性能满足既定指标要求的前提下,尽可能压缩不必要的冗余计算,提高算法效率。在算法设计中,关键门限的设置被建模为数学上的非凸优化问题,通过寻找可行域内的若干局部最优解,使得算法的计算复杂度(近似)达到最低,同时在理论上确保帧同步系统的虚警、漏警概率满足预期指标要求。仿真表明,基于非凸优化门限搜索的LDPC码辅助高效突发帧同步算法能够工作于低信噪比条件下,且具备优异的检测性能(Es/N0=-3d B时,Pf=Pm≤4x10-8);对比其它参考文献所列算法,该算法无论在虚警、漏警概率等检测性能方面,还是在计算复杂度方面,均有明显优势。
[Abstract]:In this paper, the synchronization technology in the high dynamic communication link of the low orbit satellite is studied. The main difficulties and the development trend of the modern satellite communication synchronization technology are analyzed. On the basis of this, the synchronization problem in the high dynamic communication link of the low orbit satellite is refined into three sub problems: low signal to noise ratio, large frequency offset and fast speed. Carrier frequency capture in time-varying frequency conditions; low signal to noise ratio, carrier phase tracking under fast time-varying frequency conditions; fast burst frame synchronization under low signal to noise ratio. The main research content and innovative results of this paper are as follows: (1) analysis of the frequency spectrum flattening effect of linear frequency modulation (LFM) signal and proposed LFM signal The concept of "optimal detection interval" is used to derive the theoretical expression of the best detection interval and its detection performance, and the problem of insufficient resolution of the standard discrete Chirp Fourier transform (DCFT) algorithm for carrier frequency capture in high dynamic communication links is discussed, and the improved MDCFT algorithm is proposed by introducing the fractional factor to the Chirp. In the base function, the resolution of the target area is refined. In view of the LFM signal parameter estimation problem, a MDCFT based frequency modulation rate two-dimensional space maximum likelihood search algorithm is proposed, and the computational complexity of the algorithm is analyzed. On this basis, a parameter estimation algorithm based on variable norm MDCFT (VS-MDCFT) is proposed, which will be calculated. The large amount of MDCFT is split into two smaller sub MDCFT, which reduces the computation of about 60% on the premise of ensuring that the probability of capture is not obviously reduced. In view of the parameter estimation of the nonlinear frequency modulation (NLFM) signal, the suitability of the "best detection interval" to the NLFM signal is discussed, and the optimal detection window polynomial Ch is proposed. The parameter estimation algorithm of irplet transform (ODW-PCT). The theoretical analysis and simulation show that the acquisition probability of the ODW-PCT algorithm with the best window function is obviously superior to the general PCT algorithm. (2) the signal model of the high dynamic communication link of the low orbit satellite in the phase tracking phase of the carrier phase is derived, and the general carrier phase tracking based on the Bayesian estimation theory is given. Structure. The limitations of the standard Calman filter (KF) applied to carrier phase tracking system are analyzed, and the general signal processing flow with extended Calman filter (EKF) and particle filter (PF) for carrier phase tracking is derived. From the problem of selecting the density function of the importance sampling of particle filter, the strong excitation and augmented particle set likelihood row is proposed. The sequential sampling algorithm can obtain more reliable particles which are close to the posterior probability distribution by two measures of enhancing the excitation noise and enlarging the size of the particle set. At the same time, the latest data information is introduced into the particle generation process by the likelihood ranking mechanism. The theoretical analysis and simulation results show that the strong excitation augmented particle set is based on the theory analysis and simulation results. The particle filter algorithm of likelihood ranking sampling is more stable and accurate than the Calman filter algorithm and the standard particle filter algorithm. (3) the performance effect of the system state space model on the performance of the carrier tracking system based on Bayesian estimation is analyzed, and the high order truncation model used in the general carrier tracking system is demonstrated. Based on the generation mechanism of Doppler effect, the feasibility and superiority of maneuvering target modeling technology applied to high dynamic carrier tracking system are demonstrated, and a carrier tracking algorithm based on four dimensional CS state space model is proposed, which is based on the Jerk environment. The tracking performance is improved about 1D B. compared with the general carrier tracking algorithm using the high order truncation model. The carrier tracking algorithm using a single dynamic model can not obtain the optimality energy in the high dynamic communication link with fast and time-varying relative motion state. Based on this, a carrier tracking algorithm based on multi mode fusion is proposed. The basic idea is to establish multiple parallel work sub estimators for the carrier tracking module, and the different estimator configuring different maneuvering target models. The carrier phase tracking value of the final output of the tracker is calculated by joint calculation of the output values of each sub estimator. If the number of pre assigned sub estimators is sufficient, the selected model is covered. The range is wide enough, and for the actual motion state of the communication carrier at any time, it is very likely to find a dynamic model that matches it. The carrier tracking system can always ensure the accuracy and stability of carrier phase tracking throughout the duration of the communication. (4) a burst frame synchronization system based on sliding window detection is proposed. Basic structure. For non coded auxiliary frame synchronization, the false alarm and leakage probability of frame synchronization algorithm based on hard decision and soft decision frame synchronization field (SW) detection are derived respectively. The problem of poor detection performance under low signal-to-noise ratio is discussed. A verification based on the verification of coded auxiliary frame synchronization is proposed. The LDPC code aided burst frame synchronization algorithm based on matrix detection can work at low signal to noise ratio, and its detection performance is much higher than that of non coded auxiliary frame synchronization algorithm, but there is a shortcoming that the computation complexity is too high and is not easy to be processed in real time. In order to solve this problem, a kind of LDPC code based on non convex optimization threshold search is proposed to assist the efficient burst. The basic idea of frame synchronization algorithm is to divide the frame synchronization detection into two levels, using the SW correlation and the LDPC codeword check relation to detect the frame boundary respectively. The second level detection adopts the variable operation quantity mode, and its operation amount is closely related to the first level detection measure, thus ensuring the detection performance of the two level processing algorithm. In the design of the algorithm, the setting of the key threshold is modeled as a mathematical non convex optimization problem in the design of the algorithm. By looking for some local optimal solutions in the feasible domain, the computational complexity (approximation) of the algorithm is reached to the lowest, and at the same time in theory. To ensure the false alarm of the frame synchronization system, the probability of leakage alarm meets the requirements of the expected target. The simulation shows that the LDPC code aided efficient burst frame synchronization algorithm based on the non convex optimization threshold search can work under the condition of low signal to noise ratio and has excellent detection performance (Es/N0=-3d B, Pf=Pm < 4x10-8); the algorithm is compared with other reference documents, the algorithm is compared. Both in terms of false alarm, miss detection probability and other aspects of detection performance, it has obvious advantages in terms of computational complexity.

【学位授予单位】：北京理工大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TN927.2

【参考文献】