北斗二代卫星导航接收机关键共性技术研究

【作者】 赵晋；

【导师】 张建秋；

【作者基本信息】 复旦大学 ， 电路与系统， 2010， 博士

【摘要】 全球卫星导航系统作为一种全天候、全空域的导航系统,有着其它导航系统无法比拟的优势,具有非常重要的社会经济意义和军事意义,已成为各大国空间技术竞争的焦点。作为航天大国的中国,也必然加入到这一角逐之中。北斗二代卫星导航系统是我国正在积极研究并部署的新一代卫星导航系统,具有完全的自主知识产权。投入使用后,将会为国民经济的可持续发展提供强有力的支持,同时也将在国防建设中扮演重要角色。导航接收机作为卫星导航系统的主要组成部分,是市场规模最大和产业化最核心的环节。随着北斗二代导航系统的不断发展完善,研究北斗二代卫星导航接收机关键技术,为发展我国卫星导航接收机技术的自主知识产权探索途径,具有迫切而现实的意义。由于北斗二代卫星导航系统的卫星信号标准尚未公开,本文只针对北斗二代卫星导航接收机信号捕获模块、跟踪环路模块和用户位置解算模块中的关键共性技术展开研究,从改善算法性能和降低算法复杂度两个方面寻求新的实现途径和方法。为了确定天空中的可见卫星以及卫星信号的参数,卫星导航接收机首先需要进行的工作是对卫星信号进行捕获。用户总是希望在缩短信号捕获时间的前提下尽量减少运算量。针对这一问题,本文提出一种并行差分结构的相关器用于卫星导航接收机的信号捕获。该相关器通过对相关结果的复用,在并行捕获多个卫星信号时,在运算量方面比其它已报道的改进型时域相关算法和频域相关算法具有优势,并且运算复杂度不随过采样率的提高而增加。分析和仿真结果对本文算法在运算复杂度上的优势进行了验证。对于已经捕获的卫星信号,跟踪环路负责对卫星信号进行稳定的跟踪。单比特量化的模拟前端由于结构简单,价格低廉,在低成本的卫星导航接收机中得到一定的应用。但在单比特量化接收机的载波跟踪环路中,传统反正切鉴相器性能恶化严重。针对这一现象,本文提出了一种新的鲁棒数字鉴相器,解决了在单比特量化接收机中鉴相器估计精度下降和对噪声敏感的问题。分析和仿真结果表明：在单比特量化接收机中本文提出的算法提供了比现有反正切鉴相器和数字鉴相器更好的精度和鲁棒性。针对应用更为广泛的多比特量化模拟前端,以及所在卫星导航接收机中的载波跟踪问题,本文提出一种新的基于多项式预测模型的多普勒频率跟踪方法。该方法使用多项式预测模型对多普勒频偏进行建模,用无气味卡尔曼滤波算法对其进行迭代求解,并可以实现模型的自确认,从而提供了比传统锁相环算法和其它已报道的频率跟踪算法更好的精度,适用于在中低动态及高动态卫星导航接收机中工作。卫星导航接收机的最终目的是计算出用户所处的位置。在各个卫星等效伪距误差不同分布的情况下,迭代求解用户位置时,使用加权最小二乘算法可以比最小二乘算法获得更好的性能。但在实际中等效伪距误差协方差矩阵往往是未知的。针对这一问题,本文提出一种利用迭代异方差估计算法对等效伪距误差协方差矩阵进行估计的方法,使加权最小二乘求解用户位置成为可能。仿真结果证明了本文提出算法的性能和有效性。最后,本文基于“硬件模拟前端+FPGA+PC”的结构,实现了一个满足实时性定位要求的GPS导航接收机来验证部分本文提出的方法,期待为以后北斗二代卫星导航接收机的实现提供了良好的借鉴基础。更多还原

【Abstract】 Global navigation satellite system (GNSS) is an all-weather, all-space navigation system, which has significant advantages over other navigation systems. Since its significant meaning to the community, national economy and military, GNSS has become the focus of a race in space technology among some leading countries, including China. BD-Ⅱis a new generation satellite system under development, which has been taking up with Chinese independent innovation.It will contribute to the sustainable development of national economy and play an import role in the national defense in the future.As an important part of the satellite navigation system, receiver is a key factor of navigation industrialization.Along with the development of BD-Ⅱsatellite navigation system,researches on BD-Ⅱreceivers’key technologies are very demanding. This paper focuses on common technologies in signal acquisition, tracking loop and user position calculation modules of BD-Ⅱreceiver,and presents some new methods to improve performance and reduce computation complexity.In order to determine signal parameters of visible satellites in the sky, the first task of receiver is to acquire satellite signals.Since short acquisition time with low computation complexity is preferred, a parallel differential correlation acquisition algorithm in time domain is presented in this paper, which can acquire multi-satellite signals at the same time by reusing the correlation results. We also analyze computation complexity of the proposed algorithm, comparing with traditional correlation algorithm, two modified correlation algorithms in time domain and the FFT-based correlation algorithm in frequency domain.Analysis and simulation results indicate that our algorithm has advantages in computation complexity, and its computation complexity does not increase while oversampling rate increases.After acquisition, a tracking loop is used to stably track the satellite signals.The idea of 1-bit analog-to-digital-conversion (ADC) has been widely implemented in low-cost satellite navigation receivers and low-power satellite communications due to its efficiency in bit wise processing and free from automatic gain control.The accuracy of traditional arctangent phase discriminator in tracking loop degrades greatly because of 1-bit ADC.To solve this problem,a novel phase discriminator called as robust digital phase discriminator (RDPD) is presented in this paper. Different from the existing digital phase discriminator (DPD), the proposed RDPD achieves higher precision and more robustness. Analysis and simulation results indicate the proposed RDPD has superior precision than traditional arctangent phase discriminator (APD) and DPD whenever signal noise ratio(SNR) is high or low.For more widely used multi-bits quantified front-end, we propose a novel dynamic model of Doppler frequency and phase firstly, called as a polynomial prediction model (PPM).Then a new self-validating optimal filtering method, based on PPM and the unscented Kalman filter (UKF), is developed for frequency and phase estimates in carrier tracking loop of a receiver. The analytical results show that the proposed method can effectively track the Doppler frequency and phase whatever the motion between the transceivers is. The analysis is verified by the simulations and experimental results of a real live GPS receiver, which show that the proposed method is superior to most of those reported in literatures.The final task of satellite navigation receiver is to calculate user position.When user equivalent range errors (UERE) have different distributions, the weighted least-square (WLS) algorithm has better performance than the least-square (LS) algorithm. But in practice, the covariance matrix of UERE is not known.In this paper, an iterative asymptotic unbiased heteroscedastic variance estimation method is presented, which is suitable for measurement noise with arbitrary possibility distributions. With estimation results of this method, the WLS algorithm can work in practice. Simulation results verify the performance and effectiveness of our method.At last, we develop a real-time global position system (GPS) receiver based on "font-end+FPGA+PC" architecture to verify parts of algorithms proposed in this paper, and hope this receiver to be a good design reference for BD-Ⅱreceiver in the future.更多还原