基于多传感器网络的导航系统完好性监测与信息融合研究

发布时间：2018-01-10 01:24

本文关键词：基于多传感器网络的导航系统完好性监测与信息融合研究　出处：《南京航空航天大学》2014年硕士论文　论文类型：学位论文

【摘要】：基于多传感器网络的导航系统是一种新的导航系统设计理念，将多个导航传感器系统配置在载体（如飞行器、舰船等）的不同位置，为载体的导航提供冗余的分布式测量信息的同时，还能提供计算载体电子设备局部运动补偿的惯性信息。为了提高导航系统的性能，本文对多传感器网络的完好性监测与信息融合关键技术进行研究，主要研究内容和创新点如下：（1）分析了多传感器导航系统的设计方法，给出了基于传感器网络的系统模型，并分析了多传感器数据融合结构。在此基础上着重研究了多传感器网络节点的最优化配置方法，给出了冗余传感器最优配置准则，使用平均故障时间参数对配置的可靠性进行分析，并利用单传感器可靠性贡献参数对导航系统的冗余传感器数目进行优化。由上述的研究，给出了一系列常用的传感器系统节点配置的最优化方案。（2）研究了多传感器导航系统的完好性监测方法，采用移动窗口广义似然比（MW-GLRT）完好性监测算法解决了广义似然比法（GLRT）对于特征向量提取不精确和样本奇偶向量对故障的敏感程度随时间的累积而下降的问题，显著提高了导航系统的可靠性；针对导航冗余信息不足时只能判断故障而无法隔离故障的问题，采用小波变换结合奇偶校验的综合算法，通过选取小波基和小波分解层数使检测函数对传感器故障的敏感程度区别于载体的机动，将奇偶校验函数作为故障检测隔离的保险装置，通过阶跃和斜坡故障的仿真分析验证了上述算法的有效性。（3）研究了多传感器信息融合理论，分析了常用的滤波结构。在此基础上提出了虚拟传感器及其信息融合算法，对多传感器节点降阶处理后进行信息滤波融合，仿真验证可得单节点多传感器系统通过使用该融合算法能得到高精度的相对于正交轴的测量估计值；进一步提出了一种适用于多节点导航系统的分布式信息融合算法，包括分布式系统结构、测量模型和融合算法，通过仿真验证可得，该算法充分利用了各节点的测量信息，有效提高了系统的状态估计性能。（4）建立了基于多传感器网络导航系统的数字仿真模型，包括轨迹生成、完好性监测、信息融合等模块；并设计一个半实物硬件仿真平台，并进行了初步实现。本文针对基于传感器网络的导航系统，研究了多传感器网络的完好性监测和信息融合新方法，，针对其关键技术提出了有效的解决方法，在理论方法上取得了较好的创新性研究成果，为未来高性能导航系统设计提供一定的借鉴意义和应用价值。
[Abstract]:The navigation system based on multi-sensor network is a new design concept of navigation system, which configures multiple navigation sensor systems in different positions of carriers (such as aircraft, ships, etc.). While providing redundant distributed measurement information for carrier navigation, it can also provide inertial information to calculate the local motion compensation of carrier electronic equipment, in order to improve the performance of navigation system. In this paper, the key technologies of integrity monitoring and information fusion in multi-sensor networks are studied. The main research contents and innovations are as follows: 1) the design method of multi-sensor navigation system is analyzed, and the system model based on sensor network is given. Based on the analysis of multi-sensor data fusion structure, the optimal configuration method of multi-sensor network nodes is studied, and the optimal configuration criteria of redundant sensors are given. The reliability of the configuration is analyzed by using the average fault time parameter, and the number of redundant sensors in the navigation system is optimized by using the single sensor reliability contribution parameter. This paper presents a series of commonly used optimization schemes for sensor system node configuration. 2) the integrity monitoring method of multi-sensor navigation system is studied. The generalized likelihood ratio (GML) method is used to solve the problem of MW-GLRTT (Generalized likelihood ratio) of moving window. For feature vector extraction imprecision and sample odd-even vector sensitivity to the fault decreases with the accumulation of time. The reliability of navigation system is improved significantly. Aiming at the problem that the fault can only be judged but not isolated when the redundant information of navigation is insufficient, a comprehensive algorithm of wavelet transform combined with parity check is adopted. By selecting wavelet basis and wavelet decomposition layers, the sensitivity of detection function to sensor fault is different from that of carrier, and parity check function is used as the safety device for fault detection and isolation. The effectiveness of the algorithm is verified by the simulation analysis of step and slope faults. Thirdly, the theory of multi-sensor information fusion is studied, and the common filter structure is analyzed. Based on this, a virtual sensor and its information fusion algorithm are proposed. After reducing the order of the multi-sensor nodes, the information filtering fusion is carried out, and the simulation results show that the single-node multi-sensor system can get the high accuracy relative to the quadrature axis by using the fusion algorithm. Furthermore, a distributed information fusion algorithm for multi-node navigation system is proposed, including the structure of distributed system, measurement model and fusion algorithm, which can be verified by simulation. The algorithm makes full use of the measurement information of each node and effectively improves the performance of the system state estimation. (4) A digital simulation model based on multi-sensor network navigation system is established, including track generation, integrity monitoring, information fusion and so on. The hardware simulation platform is designed and implemented. For the navigation system based on sensor network, this paper studies a new method of integrity monitoring and information fusion of multi-sensor network, and puts forward an effective solution to its key technology. Good innovative research results have been obtained in theory and method, which will provide some reference and application value for the design of high performance navigation system in the future.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN96

【参考文献】