支持向量机在初始对准状态估值中的应用

The application of SVM to state estimation of initial alignment

在捷联惯导系统初始对准状态估值过程中,针对传统Kalman滤波器的实时性问题,提出了基于支持向量机(SVM)网络的滤波器。采用闭环Kalman滤波器的输入和输出数据作为训练样本对进行训练得到SVM模型。通过仿真实验对比表明,采用SVM网络滤波器的估计精度与采用闭环Kalman滤波器的相当,但估计速度比较快。

In the state valuation process of initial alignment of strapdown inertial navigation system, a filter based on the support vector machine （ SVM ） is developed to solve the real-time problem of traditional Kalman filter. Pairs of input and output data of closed-loop Kalman filter are adopted in this paper as training samples to obtain the SVM model. Through the simulation experiment contrast, the application of SVM network filter has the same estimation accuracy when using closed-loop Kalman filter, but the estimated speed is higher.