【作者】 陈坡；

【导师】 孙付平；

【作者基本信息】 解放军信息工程大学 ， 大地测量学与测量工程， 2013， 博士

【摘要】 GNSS/INS深组合导航系统是目前研究的热点，是一种深层次的组合模型，涉及到GNSS接收机内部跟踪环路，数据信息融合的程度更加深入。其本质表现为：利用INS导航解算信息和卫星星历，计算得到载体至卫星视线方向的多普勒频移，并实时反馈至GNSS接收机跟踪环路，使GNSS接收机对动态应力不敏感；同时利用组合导航滤波器输出信息修正INS，使GNSS接收机跟踪环路仅跟踪由INS解算误差、GNSS接收机晶振、外部噪声等导致的多普勒频移误差，有效提高GNSS信号跟踪性能。当GNSS信号受干扰或遮挡时，INS导航解算信息可以实现无缝导航，同时可以预测载体的动态信息，对GNSS信号的多普勒频移和相位偏移进行连续估计，提高GNSS信号重新捕获能力。本文以GNSS/INS深组合导航理论方法为研究目标，从深组合导航系统实现的基础层面和整体性能提升层面进行研究分析。内容涵盖了GNSS/INS组合导航理论基础及仿真平台、INS辅助GNSS信号捕获与跟踪算法研究、GNSS/INS深组合导航模型算法研究、GNSS/INS深组合导航模型故障诊断算法研究。论文的主要工作归纳如下：1、构建了GNSS/INS深组合导航仿真平台。参照加拿大PLAN研究团队开发的GSNRx-utTM设计思想，基于C#语言构建了GNSS/INS深组合导航仿真平台。该仿真平台包括运动轨迹发生器、GNSS中频信号仿真器、GNSS接收机基带信号处理、IMU信号产生器、INS力学编排、组合导航滤波器等模块，可以方便进行GNSS/INS深组合导航性能评估、验证深组合不同模型及滤波算法的优劣，是后续研究的基础。2、提出了一种INS辅助GNSS接收机信号捕获算法。对GNSS信号捕获性能指标（信号检测统计指标和平均捕获时间）进行分析，着重分析研究了预检测积分时间、多普勒频移估计偏差和码相位估计偏差对捕获性能的影响；针对高动态条件下GNSS信号快速捕获的问题，从机理上进行GNSS信号捕获与INS的适配性分析，提出了一种INS辅助GNSS接收机信号捕获算法，该算法利用INS解算的位置、速度信息结合星历得到码相位和多普勒频移的先验信息，作为码/载波NCO的中心，以INS解算的位置、速度误差的不确定度作为搜索范围，缩短GNSS接收机信号捕获的时间，提高捕获性能。仿真结果表明：采用低精度的INS辅助GNSS信号捕获，在120s内，捕获灵敏度至少能提高4.2dB-Hz；在GNSS接收机冷启动模式下，信号平均捕获时间从132s降低到62s。3、研究并分析了INS辅助GNSS接收机跟踪环路结构及方法。对GNSS接收机跟踪环路DLL/PLL/FLL鉴别器进行研究；分析了环路噪声带宽、载体动态特性、接收信号载噪比对DLL/PLL/FLL跟踪误差的影响。设计了一种高动态GNSS接收机跟踪环路结构；提出了一种迭代求解带宽门限方法。重点分析了INS辅助GNSS接收机载波跟踪环路数学模型和结构方法。通过仿真分析优化了INS辅助下GNSS接收机跟踪环路参数设置，提高深组合导航系统的跟踪性能。仿真结果表明：在高动态条件下，采用低精度的INS辅助三阶PLL环路可以保持对低载噪比信号的持续锁定、跟踪状态。4、推导了集中式GNSS/INS深组合导航系统的数学模型。对深组合导航系统模型进行研究分析，详细推导了集中式GNSS/INS深组合导航系统数学模型，给出了组合导航滤波器的量测方程和状态方程；基于软件外推法来实现不同参数间的时间同步问题。对级联式和集中式深组合模型进行仿真分析，结果表明：集中式深组合导航系统在跟踪性能和导航性能方面均优于级联式深组合导航系统；中低速运动时GNSS/INS深组合导航性能整体上优于高速运动时深组合导航性能；当GNSS信号衰减时，深组合导航系统的跟踪环路误差增大，但环路仍处于锁定跟踪状态，导航精度有所降低，但仍能满足用户较高的精度要求。5、针对集中式GNSS/INS深组合导航系统量测模型的非线性问题，将UKF算法引入到集中式深组合导航滤波器中。仿真结果表明，基于UKF算法的集中式深组合导航模型在跟踪性能和导航性能方面优于基于kalman算法的集中式深组合导航模型。6、提出了一种适用于集中式GNSS/INS深组合导航系统的故障诊断与隔离方法。为避免某一通道相关器故障信息“污染”整个系统，提出了一种利用残差序列建立其时间序列模型的故障诊断与隔离方法，在GNSS接收机通道相关器输出I/Q异常信息时保证了深组合导航系统的精度。仿真结果表明：本文提出的故障诊断方法可以较好地避免故障信息对导航性能的影响。设计开发了集中式GNSS/INS深组合导航一体化软件，为后续的工程实现打下基础。更多还原

【Abstract】 GNSS/INS deeply coupled integrated navigation system has become the research hotspot inrecent years, and it is a kind of deeply integrated model, which involves the GNSS receivertracking loop, and much deeper information fusion degree. The essential performance ofGNSS/INS deeply integrated navigation system is: using INS navigation solution and satelliteephemeris, calculating the Doppler frequency in the direction from receiver to satellite, andfeedback to the GNSS receiver tracking loop in real time to make the GNSS receiver notsensitive to dynamic stress; at the same time, using the integrated navigation filter to exportcorrection message of INS, this insure the GNSS receiver only tracks Doppler error caused byINS calculation error, GNSS receiver crystal and external noise. When GNSS signal is interferedor shaded, INS navigation solution can achieve seamless navigation, predict the dynamicinformation of carrier and then make continuously estimate of Doppler of GNSS receiver andphase shift, Doppler frequency and phase shift information continuously estimated, which cangreatly reduce the signal re-capture time of GNSS receiver.This paper aims at studying the theory of GNSS/INS deeply coupled integrated navigation,the basic of realization and the overall performance promotion of deeply coupled integratednavigation system. The contents cover the GNSS/INS integrated navigation theory, simulationplatform, INS aided GNSS signal acquisition and tracking, algorithms of GNSS/INS deeplycoupled integrated navigation, fault diagnosis of GNSS/INS deeply integrated navigation model.The main work and conclusions of the paper are summarized as follows:1. The GNSS/INS deeply coupled integrated navigation simulation platform is constructed.The GSNRx-utTM design ideas of PLAN research team in Canada are referenced to develop theGNSS/INS deeply coupled integrated navigation simulation platform based on C#programlanguage. Several modules are included in the simulation platform, such as trajectory generator,GNSS intermediate frequency signal simulator, GNSS receiver baseband signal processing, IMUsignal generator, INS navigation algorithm, integrated navigation filter, which are convenient forevaluation and validation of GNSS/INS deeply coupled integrated navigation performance, themerits of the integrated filtering algorithm, and also are the foundation of the follow-up study.2. A kind of INS aided GNSS receiver signal acquisition algorithm is designed. Acquisitionperformance of GNSS signal (signal detection statistics index and the average capture time) isanalyzed, and the integration time and Doppler estimation deviation and code phase estimationdeviation are emphatically studied to evaluate the influence on performance of acquisition. Inallusion to the fast acquisition of GNSS signals under the condition of high dynamic, the arithmetic of INS aiding GNSS receiver to acquire signal was proposed to analyze the suitabilityof GNSS signal and INS. The transcendental information of code phase and Doppler frequencyshift are gotten by using the position and velocity of INS combined with ephemeris information,as the center of code/carrier NCO. INS error as the search boundary, the time of GNSS receiversignal acquisition is shorten, and acquisition performance is improved. The simulation resultsshows that the low accuracy of the INS aided GNSS signal capture, in the120s, acquisitionsensitivity can be increased by4.2dB-Hz at least; Under the cold start, the average acquisitiontime decreases from132s to4s.3. The structure and method of INS aidding GNSS receiver tracking loop are researched andanalyzed. DLL/PLL/FLL of GNSS receiver tracking loop discriminator is studied.The carrierloop noise bandwidth, dynamic characteristic and received signal carrier noise ratio to theDLL/PLL/FLL tracking error are analyzed. A high dynamic GNSS receiver tracking loopstructure is designed. A bandwidth threshold iterative algorithm is proposed. Focusing on theanalysis of INS aided mathematical model of the GNSS receiver carrier tracking loop andstructural method. Through the simulation optimization INS aided by GNSS receiver trackingloop parameter Settings, the tracking performance of deep integrated navigation system wasimproved. Simulation results show that under the condition of high dynamic, the low accuracy ofINS aided third-order PLL loop can keep locking and tracking signal with low noise ratio.4. The centralized deep GNSS/INS integrated navigation system mathematical model isdeduced. System model of centralized deep combination is analyzed. Centralized deepGNSS/INS integrated navigation system mathematical model is deduced in detail. Themeasurement equation and status equation of the integrated navigation filter are given. Theextrapolation method based on software is used to achieve time synchronization problembetween different parameters. The cascading and centralized deep combination model aresimulated and analysed, the result shows that centralized deep integrated navigation system ontracking performance and the navigation performance is better than cascading deep integratednavigation system under the condition of low dynamic and high dynamic. when The performanceof the deep GNSS/INS integrated navigation is better in slow speed motion overall high speed.When GNSS signal attenuation, the tracking loop error of deep integrated navigation systemincreases, but as the loop is still locked, the navigation precision turns lower, but can also meetthe users’ requirement of high precision.5. Using UKF in the filter to solve the nonlinear problem of the centralized GNSS/INSintegrated navigation. Using UKF algorithms in integrated navigation filter. Simulation resultshows that the tracking performance of centralized deeply integrated navigation model based onUKF is better than that of kalman filter. 6. A kind of fault diagnosis and isolation method is designed for the centralized GNSS/INSintegrated navigation system. For the sake of avoiding one certain channels correlator faultinformation "polluting" the whole system, the method using residual error sequence to establishthe time series model of the GNSS receiver correlator was proposed. At the time of GNSSoutputting I/Q channels abnormal information this guaranteed the precision of the integratednavigation system. Simulation results show that the proposed fault diagnosis method is a goodway to avoid the fault information’s influence on navigation performance. A centralizedGNSS/INS integrated navigation deep integration software was designed and developed, thismade a firm foundation for the subsequent engineering realization.更多还原