基于BP神经网络组合导航融合算法

Integrated Navigation Fusion Algorithm Based on BP Neural Network

文章针对卡尔曼滤波算法的缺陷,建立了组合导航系统模型,提出了基于BP神经网络修正卡尔曼量噪权值的自适应卡尔曼滤波法,对量测噪声统计特性进行了自适应调整,并与传统卡尔曼算法进行计算机仿真比较。结果表明,基于BP神经网络的卡尔曼滤波器在一定程度上抑制了数据的发散,提高了导航精度。

The adaptive kalman filter based on BP neural network to correct weights of measurement noise is proposed via the analysis of limitation of kalman filter. The statistical characteristic of measurement noise is adaptive adjusted after the a integrated navigation model is set up. Comparing the traditional kalman filter, the simulation result show that the new algorithm prevent divergence effectively. The precision of navigation is improved.