旋转状态下GNSS信号跟踪算法研究及GPU实现

发布时间：2018-09-14 09:12

【摘要】：GNSS(Global Navigation Satellites System,全球卫星导航系统)由于其成本低廉、长时间导航精度高等特点,在军事领域有着广泛的应用。然而由于许多武器存在旋转,旋转会引起信号多普勒的周期性变化,对信号正常接收产生一定影响,一般的接收机在这种情况下很难保持信号的稳定跟踪。因此,分析旋转状态下锁相环的跟踪性能以优化环路参数、设计合理的算法至关重要。本文以分析旋转状态下环路的跟踪性能和参数优化为目标,主要做了以下工作:旋转状态下GNSS接收信号的幅度、多普勒和相位等均会发生复杂的变化,本文根据物理学运动定律和入射角相对天线的变化,结合天线的方向图,对旋转状态下的接收信号进行精确建模,指出旋转状态下接收信号的多普勒呈正弦规律变化。旋转对载波多普勒的影响较大,对伪码多普勒的影响较小,分析时可不用考虑旋转对码跟踪的影响。旋转引起多普勒和相位的正弦变化,导致跟踪环的跟踪误差不能收敛到一个固定的值,传统的采用计算动态应力的分析方法不再适用。本文以锁相环误差传递函数为工具,建立了输入相位幅度和相位跟踪误差幅度之间的关系,给出了一种旋转状态下锁相环性能分析的方法,并对该方法做了仿真,验证了其合理性与可行性。利用本文提出的分析方法,数值分析了旋转状态下锁相环、锁频环在不同噪声带宽下的跟踪灵敏度以及在不同载噪比下能稳定跟踪的最大转速,数值仿真给出采用锁相环跟踪信号时的最佳噪声带宽,为旋转状态下环路的优化设计提供了理论支撑。比较了不同载波跟踪环路进行信号跟踪的优缺点,指出噪声带宽相同时,二阶锁相环和三阶锁相环跟踪性能的优劣与转速密切相关,在低转速下选择二阶环,在高转速下选择三阶环。GNSS软件接收机的实时处理是其重要的研究方向,本文借助于GPU强大的并行处理能力,给出了一种基于CPU+GPU的实时软件接收机总体架构,并对信号跟踪环节做了详细的设计。采用该软件接收机对实测的旋转信号处理结果验证了文中模型的正确性和分析方法的合理性,对于采样率为5MHz的仿真信号,同时跟踪12颗卫星,每处理1ms信号的平均耗时仅为0.109ms。
[Abstract]:GNSS (Global Navigation Satellites System, (Global Satellite Navigation system) is widely used in military field because of its low cost and high precision. However, due to the rotation of many weapons, rotation will cause periodic changes of the signal Doppler, which has a certain impact on the normal reception of the signal. In this case, it is difficult for the general receiver to keep the stable tracking of the signal. Therefore, it is very important to analyze the tracking performance of PLL in rotation to optimize the loop parameters and to design a reasonable algorithm. The aim of this paper is to analyze the tracking performance and parameter optimization of the loop under the rotating state. The main work is as follows: the amplitude, Doppler and phase of the received signal of the GNSS under the rotating state will change in a complex way. According to the law of motion of physics and the change of incidence angle relative antenna, combining the pattern of antenna, the received signal under rotation state is modeled accurately, and it is pointed out that the Doppler of received signal in rotating state changes in sinusoidal law. The effect of rotation on carrier Doppler is greater than that on pseudo code Doppler, so it is unnecessary to consider the effect of rotation on code tracking. Due to the sinusoidal variation of Doppler and phase caused by rotation, the tracking error of the tracking loop can not converge to a fixed value, so the traditional analysis method to calculate the dynamic stress is no longer applicable. Using the error transfer function of PLL as a tool, the relationship between the input phase amplitude and the phase-tracking error amplitude is established in this paper. A method for analyzing the performance of PLL in rotating state is presented, and the simulation of the method is given. The rationality and feasibility are verified. By using the analytical method presented in this paper, the tracking sensitivity of phase-locked loop in rotating state, the tracking sensitivity of frequency locked loop under different noise bandwidth and the maximum speed of stable tracking under different load / noise ratio are numerically analyzed. The optimal noise bandwidth of PLL tracking signal is given by numerical simulation, which provides theoretical support for the optimal design of loop under rotating state. The advantages and disadvantages of different carrier tracking loops for signal tracking are compared. It is pointed out that when the noise bandwidth is the same, the performance of second-order PLL and third-order PLL is closely related to the rotational speed, and the second-order loop is chosen at low speed. It is an important research direction to select the real-time processing of third-order ring. GNSS software receiver at high speed. This paper presents an overall architecture of real-time software receiver based on CPU GPU with the help of the powerful parallel processing ability of GPU. The signal tracking link is designed in detail. The software receiver is used to process the measured rotational signals to verify the correctness of the model and the reasonableness of the analytical method. For the simulated signals with 5MHz sampling rate and tracking 12 satellites at the same time, the average processing time of each 1ms signal is only 0.109 Ms.
【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN967.1

【相似文献】