【作者】 张爱军；

【导师】 王昌明；

【作者基本信息】 南京理工大学 ， 测试计量技术及仪器， 2009， 博士

【摘要】 高精度的导航定位不仅是水下潜器获取有效信息的必要条件,而且决定了它是否可以安全作业及返回。因此,高精度导航定位是研究水下潜器的关键技术之一。由于水成为一个有效的法拉第笼,所以许多基于电磁传播的导航方法(如GPS等)在水下不易实现。虽然船位推算(DR)和捷联惯性组合系统在短期内具有较高的定位精度,但定位误差会随时间发散。融合GPS、船位推算和捷联惯性组合等信息有利于提高导航定位系统的精度、可靠性和适应性,是水下导航定位的一个发展趋势。论文根据实际需求,设计了利用捷联惯性组合系统作为主系统,多普勒计程仪、GPS、电子罗盘和由螺旋桨转换过来的速率信息构成的DR系统作为辅助子系统的水下组合导航定位系统。结合采用的导航传感器,建立组合导航数学模型。由于标准卡尔曼滤波不能处理非线性模型,而扩展卡尔曼滤波有计算量大等缺点,故采用无迹卡尔曼滤波(UKF)进行处理。并且,在水下组合导航系统中,由于系统本身元器件(如陀螺和加速度计等)的不稳定性以及外部应用环境不确定因素的影响,系统噪声和观测噪声统计特性的准确描述非常困难,同时也不能保证UKF滤波的收敛性和稳定性。因此,本文中提出一种自适应UKF算法,仿真验证了算法的可行性。同时由于联邦卡尔曼滤波具有更好的容错性和设计灵活性,选择联邦卡尔曼滤波作为组合导航的基本融合算法。为使船位推算导航法能客观地反映实际航向情况,从而提高船位推算精度,不仅要发展新的测向仪器和测速仪器,而且要研究推算误差补偿方法。本文在建立船位推算误差模型基础上,对不同误差源分别提出误差补偿方法,最后进行了实验验证。初始对准误差是捷联式惯导系统的主要误差源之一,它直接影响到惯导系统的工作精度。论文推导了静基座对准误差模型,并进行可观测性分析,然后进行试验研究。由于在捷联惯组的实际应用过程当中,无论是机载、舰载、车载等,都会受到周围环境的影响(特别是水中摇摆情况下),因此有必要对动基座对准技术进行研究。如何在较短时间和恶劣的环境下进行高精度的初始对准,对于提高水下组合导航系统机动性有着极其重要的意义。本文对动基座对准技术进行了初步研究。更多还原

【Abstract】 The high precision integrated navigation system(INS) is not only the necessary condition of obtaining effective information about Underwater Vehicle,but also the determining whether it can safely work and return or not.So,the high precision navigation and position technology is one of the key technologies of underwater vehicle.Because the Faraday cage formed by water,many navigation methods based on electromagnetism transiting(such as GPS) are difficult to be realized underwater.The position precision of Dead Reckoning(DR) and Strap-down Inertial Navigation System(SINS) is higher in little time,but the error will be more and more emanative along with time.The information amalgamation of GPS,DR,and SINS is of benefit to increase the precision,reliability,and applicability,which is a developing direction.According to the practical demands,an underwater INS is designed of which the main system is SINS and the accessorial system is DR.DR consists of Doppler Velocity Log (DVL),Electronic Compass(EC),GPS and the speed which is transformed from airscrew. The nonlinear mathematic model of INS is established with the help of the used sensors.The Standard Kalman Filter can not solve the nonlinear mathematic model,and the Extending Kalman Filter(EKF) will consume much more calculation,so Unscented Kalman Filter(UKF) is adopted in this system.Due to the instability of sensors and the uncertainty of outer environment factors,it’s difficult to analyze the statistical characteristic of system and observational noise exactly and to ensure the convergence and stability of UKF.Then a self-adaptive UKF is advanced in this paper,and the algorithm is proved by simulation. Federated Kalman Filter will be chosen to be a basic fusion algorithm of INS,which is good at error-allow and design convenience.In order to show the actual course and enhance the precision of DR,not only the new type angle and velocity measurement instruments should be developed,but also the estimate error compensation methods should be researched.The corresponding error compensation methods aiming at different error origins are put forward through the establishing of the error model of DR,and the methods are proved by test.The initial alignment error is one of the main errors of SINS,and it affects the precision directly.An error model of initial alignment with stationary base is advanced,the observation analysis is done,and the corresponding test research is achieved.It is necessary to research the initial alignment technology with moving base because of the environment effects (especially the water waving) on plane-carried,ship-carried,or vehicle-carried during the practical operating.How to quickly achieve a higher precision initial alignment in an abominable environment has a very important significance to enhance the flexibility of INS, and which is researched in the thesis.更多还原