基于RBF网络的惯导系统初始对准

Initial Alignment of Inertial Navigation System Based on the RBF Neural Network

建立了惯导系统（ＩＮＳ）初始对准的线性和非线性误差模型。分析了径向基函数（ＲＢＦ）网络的结构和工作原理。研究了ＲＢＦ网代替初始对准中的卡尔曼滤波器的实现方法。通过仿真表明，用神经网络进行初始对准，既可获得与卡尔曼滤波相同的对准精度，又提高了系统的实时性。

Linear and nonlinear error models for Inertial Navigation System( INS ) initial alignment are established. The structure and principle of Radial Basis Function ( RBF ) neural network are studied. A method of RBF neural network instead of Kalman filter in the initial alignment is introduced. The simulation results show that using neural network in initial alignment can obtain alignment accuracy which is similar to that of the Kalman filter. In the meantime the alignment time is reduced considerably.