摘要
为了提高自主水下航行器(AUV)组合导航系统精度,选择了捷联式惯性导航系统(SINS)、多普勒速度声纳(DVS)以及地形匹配导航系统(TAN)作为AUV组合导航系统导航传感器,建立了AUV组合导航系统的状态模型和导航传感器观测模型,运用了一种基于径向基函数(RBF)神经网络进行H∞滤波信息分配的信息融合方法,并进行了计算机软件仿真。仿真结果表明,在有色噪声情况下,AUV组合导航系统的导航姿态、速度和位置精度得到了提高,有效地克服了传统滤波容易发散的缺点,提高了AUV组合导航系统的容错性能和导航精度。
To improve the autonomous underwater vehicle （AUV） navigation accuracy, strap-down inertial navigation system （SINS）, Doppler velocity sonar （DVS） and terrain aided navigation （TAN） were adopted in the AUV integrated navigation system. Mathematical models of the AUV integrated navigation system and an observation model of the chosen navigation sensors were built according to the system simulation data. An improved filter based on radial basis function （RBF） neural network for adjusting the information sharing factors was designed and implemented in the AUV integrated navigation system. Simulation resuits show that the navigation accuracy is improved obviously with the specified sensors and H∞ filter in the ease of colored noise. The novel integrated navigation system can effectively suppress the divergence of the filter, improve the fault tolerance ability, and greatly raise the navigation accuracy.
出处
《鱼雷技术》
2009年第1期14-17,共4页
Torpedo Technology
基金
国防科技重点实验室基金资助项目
