【作者】 张涛；

【导师】 袁赣南；

【作者基本信息】 哈尔滨工程大学 ， 导航、制导与控制， 2010， 博士

【摘要】 GPS/SINS超紧密组合导航系统增强了对GPS的修正,利用组合导航滤波器估计的导航信息辅助GPS的信号跟踪过程,增强了GPS的抗干扰能力,增强了组合系统在弱信号和高动态环境下的导航性能。本文主要对GPS/SINS超紧密组合导航系统的GPS修正、可观测性分析、辅助信息异常检测和导航数据同步、组合导航滤波这些关键问题进行了研究和探讨。目前超紧密组合导航系统GPS修正方法的组合导航滤波器计算量大,系统可靠性差,GPS跟踪过程容易受到不稳定跟踪环路的影响,针对这些问题本文提出了双跟踪闭环超紧密组合导航系统GPS修正方法。双跟踪闭环修正方法通过增强GPS跟踪环的跟踪性能来提高GPS的性能,GPS的跟踪过程以自主跟踪为主,由SINS输出的导航信息对跟踪信号的多普勒频移进行估计,提供对GPS跟踪过程的外部修正。跟踪环内码与载波跟踪一同进行,相互辅助,利用码跟踪较好的抗干扰和载波跟踪的高精度特性,增强了GPS的自主跟踪性能。组合导航滤波器测量GPS与SINS的导航偏差,实时的估计、修正SINS的误差,修正后的SINS为GPS跟踪过程提供了高更新率和可靠的修正信息,增强了跟踪过程的抗干扰能力。由于各通道的跟踪过程没有直接的联系,减小了跟踪过程的相互干扰,增强了在弱信号环境下的跟踪性能。载波频率误差修正与载波相位误差修正分离,分别由通道滤波器和组合导航滤波器估计、修正,减小了组合导航滤波器的计算量。仿真结果表明,双跟踪闭环修正方法增强了GPS跟踪过程在弱信号和载体高动态运动环境下的性能,满足提高超紧密组合导航系统性能的要求。可观测性分析是超紧密组合导航系统设计必不可少的一部分。由于超紧密组合导航系统复杂的时变特性,应用一些传统的可观测分析方法对其可观测性进行分析较为困难。本文通过对超紧密组合导航系统的分析,从可观测性的定义出发,根据系统的特点提出了一种超紧密组合导航系统可观测性分析的方法。根据系统的特性,通过一系列降维和近似处理简化了可观测性的分析过程,将载体的运动分解,通过对分解运动的可观测性分析来了解组合系统的可观测性随载体运动变化的情况,为系统的设计提供了指导。提出了一种基于局域波分析的多普勒频移异常检测方法。针对经验模式分解方法筛选条件不严格和分解存在伪分量的问题,改进了筛选条件,提出了伪分量的消除方法。利用多普勒频移信号所具有的自相似特性,应用局域波方法对多普勒频移与同相信号在相同时刻自相似性的变化进行分析,得出多普勒频移异常的判断。实验分析结果表明,此方法可以有效的检测出多普勒频移所出现的异常。以AR模型估计为基础,提出了SINS辅助信息时序估计方法,有效地抑制了短时SINS修正扰动。提出了以GPS秒输出脉冲为时间基准的GPS与SINS的时间同步方法。针对通道内滤波过程、组合导航滤波过程的观测时刻与滤波时刻不同步问题,提出了数据同步方法,根据数据更新同步时间差,分别由通道滤波器和导航滤波器所得到的最优估计值,利用时间更新和状态更新的方法对滤波时刻的观测值进行估计,减小滤波误差。仿真实验表明,本文所提出的数据同步方法有效的减小了时序误差,增强了GPS跟踪过程和组合导航滤波过程的精度。组合导航参数估计是组合导航系统研究的最关键问题。针对滤波过程因误差模型不正确和噪声估计不准确导致滤波估计误差较大的问题,通过对超紧密组合导航系统的研究,提出了以最大似然估计为基础的系统噪声和量测噪声的估计方法。对超紧密组合导航系统的可观测性分析表明,系统中存在不可观测量,且随着载体的运动状态的不同可观测性将发生变化,状态的不可观测持续时间一般较短。状态可观测性的改变将对组合导航滤波过程产生很大影响,导致滤波误差增大甚至发散。本文提出了以模糊控制规则为基础的渐消记忆自适应卡尔曼滤波方法。该方法对卡尔曼滤波的增益矩阵设置了权值,通过改变权值来调整观测信息对新状态估计所产生的影响,由前面对超紧密组合导航系统可观性的分析结果,设计了权值的模糊调整方法。由载体的速度、加速度和姿态测量信息,由已得到的可观测性分析结果,实时调整权值的大小,减小不可观测状态对滤波过程的影响。仿真实验结果表明,模糊渐消记忆自适应卡尔曼滤波方法减小了状态可观测性短暂变化的影响,增强了滤波的精度,提高了超紧密组合导航系统的性能。更多还原

【Abstract】 Ultra-tightly integrated GPS/SINS navigation system enhances the GPS correction, use the estimation of the navigation data to aid the process of GPS signal tracking, which enhances the anti-jamming capability and the navigation performance of the integrated system in the weak signal and high-dynamic environment.The filter has large amount of calculation with GPS corrected algorithm for ultra-tightly integrated GPS/SINS navigation system, GPS tracking process vulnerable to the impact of unstable tracking loop. To address these issues, this paper proposes a dual-track and closed-loop GPS correction for ultra-tightly integrated GPS/SINS navigation system. The dual-track and closed-loop correction algorithm improve GPS performance by enhancing the tracking performance of GPS tracking loop. GPS mainly track the signal independently, from the SINS output navigation information to estimate the Doppler frequency of the tracking signal, providing GPS tracking process with external amendments. Within the tracking loop, the code and carrier track together for mutual support. Use a good anti-interference features of code tracking and high-precision features of carrier tracking, which enhanced GPS autonomous tracking performance. The integration navigation filter measurements with the SINS navigation deviation, real-time estimate and correct the SINS errors, corrected SINS provide high update rate and reliable information to enhance the anti-jamming capability for GPS tracking. As a result of channels of the tracking loop are not directly connect, which reduce the mutual interference of the tracking process and enhance the tracking performance in weak signal. The correction of the carrier frequency and phase tracking error are separated, the channel filter and integrated filter estimate and correct the errors, which reduce the load. The simulation results show that the dual-track and closed-loop corrected algorithm enhance the performance of the GPS tracking process in weak signal and high dynamic environments, which meet the requirements which increased the performance of the ultra-tightly integrated navigation system.Observability analysis is important for ultra-tightly integrated navigation system design. Use the traditional method to analysis the observability of the ultra-tightly integrated navigation system is difficult, because of the complex time varing characteristics. Form the definition of the observability, this paper proposes a observability analysis method for ultra-tightly integrated navigation system by analyzing the ultra-tight integrated navigation system. According to the characteristics of the system, through a series of approximation, which simplifies the process of observability analysis. Through the decomposition of the movement and the observability results understand the information that the observability variation with the body movement, which provide guidance for system design.A detection method of abnormal Doppler frequency based on local wave analysis is proposed. According to the problem that the empirical mode decomposition method are not strict and existing pseudo-component, improves selection conditions, proposes the elimination of pseudo-component method. Using Doppler frequency shift signal with the self-similar characteristics, analyzes the changes of self-similarity in the same time between Doppler frequency and in-phase signal with local wave method, obtains judgments of abnormal Doppler. Experimental results show that this method can effectively detect the Doppler shift arising from abnormal. Based on the AR model, this paper proposes SINS supporting information of timing estimation method, effectively suppress the short-term disturbance. A time synchronization method between GPS and SINS is proposed based on the GPS second pulse. According to the problem that observation time data and update synchronization time are different by channel filtering process and integrated navigation filtering process, this paper proposes a data synchronization method. Use the time and data update method to estimate the observations in filtering time by the optimal estimation of the channel filter and integrated navigation filter output, reducing the filtering error. The simulation results show that the algorithm proposed by this paper effectively reduce the time warp and enhance the accuracy of GPS tracking process and integration filtering process.Navigation parameter estimation is most important issue for integrated system. According to the problem which incorrect error model and inaccurate noise estimation lead to larger filter error. Through study the ultra-tightly integrated navigation system, the estimation method of system noise and measurement noise based on the maximum likelihood estimation is proposed. Analyzing the observability of the ultra-tightly integrated navigation system, which results show that there have unobservable state within system, this state change with the movement of body , this unobservability generally much shorter. The variation of state observability have a significant impact to integrated navigation system, leading to increased filtering error or divergence. This paper proposes a adaptive fading memory kalman filtering method based on fuzzy rules. This method sets a weight for kalman filter gain matrix, by changing the weight to adjust the impact from observable information to estimate the new state. Based on the observability analysis results for ultra-tightly integrated navigation system in front, designing the method for adjusting the weights. From the velocity, accelerate, attitude information and observability analysis results, adjusting the weight in real time, which reduce the impact from unobservable state to filtering process. The simulation results show that the fuzzy fading memory adaptive kalman filtering method reduce the impact of short-time change of state observability, enhance the accuracy of filtering and performance of ultra-tightly integrated navigation system.更多还原