高精度导航接收机的群时延建模、测量和校准技术

发布时间：2018-01-12 23:07

本文关键词：高精度导航接收机的群时延建模、测量和校准技术　出处：《国防科学技术大学》2014年博士论文　论文类型：学位论文

【摘要】：高精度导航接收机广泛分布应用于卫星导航系统的地面段、空间段、控制段,是卫星导航系统完成星地时间同步、卫星精密定轨等重要系统业务的核心测量设备,是保障导航系统服务性能的基础。在高精度导航接收机中,通道群时延的非理想特性引起的信号相关峰畸变,会导致测距偏差,成为伪距测量中不可建模误差的主要来源。因此,对群时延建模、分析其对高精度测距的影响并研究其测量和校准技术很有必要。群时延与信号相关峰畸变关系密切,目前国内外普遍认为群时延只要存在波动就会产生相关峰畸变,而且特性不一样的群时延将产生不一样的相关峰畸变。但实际中会存在这种情况:两个特性差异较大的群时延可能产生基本一致的相关峰畸变,引起的时延估计不一致性很小。针对这一情况,本文提出了等效群时延的定义。该定义可简化群时延非理想特性对相关峰畸变影响的分析,并指导群时延测量和校准的设计,在进行群时延测量和校准时,只需要校准后的群时延与理想群时延等效即可。目前,一般通过仿真的方法分析群时延非理想对高精度测量的影响,不具备一般性,且效率较低。本文通过分析群时延非理想造成信号相关峰畸变的内在机理,建立了基于傅立叶级数分解的群时延通用分析模型,该模型可将任意群时延特性分解为有限阶数的余弦型群时延和正弦型群时延的级联,余弦型群时延传输函数和正弦型群时延传输函数可分解为有限阶数的线性相位传输函数的叠加,首次给出了群时延非理想与信号相关峰畸变之间的解析关系。并由此得出群时延分解的低频分量是影响信号相关峰以及测距的主要因素,高频分量影响比较小甚至可忽略。此模型有助于分析群时延非理想对时延估计的影响,并可指导群时延的测量和校准,在进行群时延测量校准时,只需要校准会导致相关峰畸变的傅立叶分解项即可。针对传统群时延测量方法存在的测量精度与测量分辨率无法兼顾的问题,论文提出了基于窄带扩频信号的群时延测量方法,该方法可以在任意分辨率下实现群时延的高精度测量,测量精度可达到0.01ns。论文同时给出了该测量方法的参数优化设计方案:早迟相关间隔越宽,群时延测量系统偏差越小;群时延波动越剧烈,码率要求越低,对于导航应用中大部分滤波器,采用1.023MHz的码率即可满足要求。在论文的仿真实例中,对波动幅度为50ns,波动周期为10.23MHz的余弦型群时延进行测量时,采用1.023MHz的码率即可达到0.01ns的精度。根据基于傅立叶级数分解的群时延通用分析模型,并结合等效群时延的定义,论文提出了一种等效群时延的高精度测量方法,该方法通过估计群时延的傅立叶分解系数来实现等效群时延的测量。仿真表明对采用该方法测量得到的群时延进行补偿后,信号相关峰与理想相关峰的形状基本一致,由群时延非理想引入的时延估计误差在0.02ns之内,可忽略不计。在传统的群时延校准方法中,模拟域校准一般存在精度与灵活性差的问题;数字域的频域校准需要做大数据量的FFT和IFFT运算,运算复杂度高;数字域的时域校准虽然复杂度低,但目前的算法无法得到时域滤波器的解析表达式,因此很难确定阶数并从理论上说明校准性能。论文提出了基于傅立叶级数模型的群时延校准算法,该算法的时域滤波器存在可解析的表达式,在理论上可以消除群时延非理想对时延估计的影响,且实现效率高,需要的滤波器阶数低。实验结果表明仅用13阶的FIR滤波器即可消除群时延波动幅度达80ns的中频滤波器引起的信号相关峰畸变和时延估计偏差。最后,对本文研究成果进行了总结,论文的主要成果均已在我国自主卫星导航系统的建设中得到应用。
[Abstract]:The control section of the ground segment, high precision navigation receiver widely applied to satellite navigation system, the space segment, is a satellite navigation system to complete the synchronous satellite time, core measuring equipment of satellite orbit determination system and other important business, is a foundation to ensure the service performance of navigation system in high precision navigation receiver, signal correlation peak distortion caused by the non ideal characteristics of the channel group delay, will lead to become the main source of ranging bias, pseudorange measurement can not be modeling error. Therefore, the group delay modeling, analysis of its impact on high precision and it is necessary to study the measurement and calibration technology. The group delay and signal correlation peak distortion close relationship at home and abroad generally believe that as long as there are fluctuations in group delay will produce peak distortion and group delay characteristics of different will have a correlation peak distortion is not the same. But in practice there will be this A: two characteristics of different group delay may have correlation peak distortion is substantially uniform, caused by the inconsistency of time delay estimation is very small. In view of this situation, this paper puts forward the definition of equivalent group delay. Analysis of the definition of group delay can simplify the influence of non ideal characteristics of the correlation peak distortion, and guide the design group delay measurement and calibration, the group delay measurement and calibration, only after calibration of the group delay and group delay equivalent to ideal. At present, generally by the method of simulation analysis on the influence of non ideal group delay measurement accuracy, does not have generality, and low efficiency. In this paper, through the analysis of internal mechanism the ideal signal correlation peak distortion caused by non group delay, group delay analysis model based on general Fu Liye series decomposition is established, the model can be arbitrary group delay characteristic is decomposed into finite order cosine type group Cascaded delay and sinusoidal group delay, group delay superposition cosine transfer function and sinusoidal group delay transfer function decomposition of linear phase transfer function is of finite order number, first given group delay non analytical relationship between the ideal and the signal correlation peak distortion. And thus the low frequency component of group delay decomposition is main factors affecting the signal correlation peak and range, high frequency components of relatively small impact or even negligible. This model is helpful to the analysis of non ideal effects on the group delay time delay estimation, and can guide the measurement and calibration of group delay, group delay measurement in the calibration, only need calibration will lead to Fu Liye correlation peak distortion decomposition you can. In the traditional methods of measuring the group delay measurement accuracy and measurement resolution can not take into account the problem, proposed group delay measurement method of narrowband signal based on the spread spectrum. Method can achieve high precision measurement of group delay at any resolution, measurement accuracy can reach 0.01ns. at the same time, given the parameter optimization design of the measurement methods: early late correlation interval is wide, group delay measurement system bias is small; group delay fluctuation is more intense, more low rate code requirements, for most navigation applications in the 1.023MHz rate to meet the requirements. In the simulation example, the volatility is 50ns, 10.23MHz is the cosine wave cycle group delay measurement, the 1.023MHz rate can reach 0.01ns accuracy. According to the analysis model based on universal group delay Fu Liye series decomposition, combined with the definition of equivalent group delay the paper presents a high precision measurement method for equivalent group delay, the method by measuring the Fu Liye group delay decomposition coefficient to achieve equivalent group delay estimation . simulation results show that the compensation of the group delay measured by this method, consistent signal correlation peak and ideal correlation peak shape, by the introduction of non ideal group delay time delay estimation error within 0.02ns, can be neglected. In the traditional group delay calibration method, analog domain calibration accuracy and flexibility are the problem of the poor; frequency domain digital calibration needs to be bigger and the amount of data FFT and IFFT computing, high computational complexity; time domain calibration of the digital domain while the complexity is low, but the current algorithm cannot get an analytical time-domain filter, it is difficult to determine the order and calibration performance in theory is put forward. This paper group time delay calibration algorithm based on Fu Liye series model, the algorithm of time domain filter expressions can be resolved, can eliminate the influence of non ideal group delay time delay estimation in theory, and the realization of High efficiency, low requirement for the order of the filter. The experimental results show that the group delay filter to eliminate the cause of fluctuation of 80ns can only use 13 order FIR filter the signal correlation peak distortion and time delay estimation error. Finally, the research results in this paper are summarized, the main results are in the construction of our own satellite navigation system application.

【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN965.5

【相似文献】