【作者】 张红星；

【导师】 张国成； 沈明学；

【作者基本信息】 哈尔滨工程大学 ， 船舶与海洋工程（专业学位）， 2022， 硕士

【摘要】 多AUV(Autonomous Underwater Vehicle)协同作业是未来水下作业的发展趋势,协同导航是其关键技术之一,主从式AUV协同导航系统不仅可以提高从AUV定位精度,而且能大大降低协同作业系统的成本,主从式AUV协同导航技术渐成为导航领域研究热点。本文建立了主从式AUV协同导航系统的状态模型和观测模型,给出常规非线性协同导航滤波流程,对协同导航系统可观测性作出分析,比较了不同滤波方法在量测噪声为高斯噪声时的性能,为后文作出铺垫。针对主从式AUV协同导航中量测信息受到水下复杂且未知的环境干扰而出现异常的问题,设计对异常量测噪声具有鲁棒性的协同导航算法,其中针对单主AUV协同导航系统出现量测野值以及量测噪声呈非高斯分布的问题,将最大互相关熵的思想引入容积卡尔曼滤波,提出最大熵容积卡尔曼滤波算法,提高了单主AUV协同导航系统定位精度。针对时变异常观测噪声,在最大熵容积卡尔曼滤波基础上,结合强跟踪滤波的思想,提出最大熵自适应滤波,引入滤波收敛判据,当滤波发散时,采用强跟踪滤波保证滤波的稳定性,当滤波收敛时,采用最大熵容卡尔曼滤波保证状态估计精度,提高了单主AUV协同导航系统的稳定性。考虑到双主AUV协同导航系统中从AUV在同一时刻接收到的两个主AUV的量测信息受到不同程度的干扰,使得来自不同主AUV的量测信息存在较大的差异,导致从AUV导航精度降低甚至发散的问题,提出一种基于信息熵最优权重分配的协同导航方法,提高协同导航的精度和系统鲁棒性。更多还原

【Abstract】 Multi-AUV collaborative operation is the development trend of underwater operations in the future,and collaborative navigation is one of its key technologies.The master-slave AUV collaborative navigation system can not only improve the positioning accuracy of the slave AUV,but also greatly reduce the cost of the collaborative operation system.AUV collaborative navigation technology has gradually become a research hotspot in the field of navigation.This paper establishes the state model and observation model of the master-slave AUV collaborative navigation system,presents the conventional nonlinear collaborative navigation filtering process,analyzes the observability of the collaborative navigation system,and compares different filtering methods when the measured noise is Gaussian noise The performance,pave the way for the following article.Aiming at the problem that the measurement information in the master-slave AUV collaborative navigation is disturbed by the underwater complex and unknown environment,the collaborative navigation algorithm is designed to be robust to the abnormal measurement noise,and the single-master AUV collaborative navigation system appears To solve the problem of non-Gaussian distribution of measurement outliers and measurement noise,the idea of maximum cross-correlation entropy is introduced into the volume Kalman filter,and the maximum entropy volume Kalman filter algorithm is proposed to improve the accuracy and robustness of the multi-AUV coordinated navigation system.Aiming at the time-varying abnormal observation noise,based on the maximum entropy volume Kalman filter,combined with the idea of strong tracking filtering,the maximum entropy adaptive filtering is proposed,and the filter convergence criterion is introduced.When the filtering diverges,strong tracking filtering is used to ensure the stability of the filtering.When the filtering converges,the maximum entropy-capacity Kalman filter is used to ensure the accuracy of the state estimation and improve the stability of the system.Taking into account that the measurement information of the two main AUVs received from the AUV at the same time in the dual-main AUV coordinated navigation system is interfered to different degrees,the measurement information from different main AUVs is quite different,which leads to the navigation from the AUV For the problem of accuracy reduction or even divergence,a collaborative navigation method based on the optimal weight distribution of information entropy is proposed to improve the accuracy of collaborative navigation and the robustness of the system.更多还原