光纤陀螺捷联系统振动误差补偿研究

Research on vibration error compensation of FOG SINS

针对舰船振动环境对光纤陀螺捷联系统性能的影响,提出基于神经网络的非线性数学补偿方案。对加速度计信息作傅里叶分析,进行振动特征提取;利用提取的振动特征值作为输入向量进行BP神经网络训练,建立光纤陀螺误差输出模型;对陀螺进行在线补偿,减小振动引起的漂移误差,提高了捷联系统初始对准精度。

To reduce the impact on characteristic of fiber-optic gym （FOG） strapdown inertial system（SINS） from vibration of vessel, a new nonlinear mathematical compensation scheme based on neural networks is proposed. Making Fourier analysis to the information on accelerometers,the vibration feature is extracted. Using the vibration feature value as input vector to train BP neural networks, the output error model of FOG is set up. Compensating the FOG online to reduce the drift error from vibration, the accuracy of alignment of SINS is enhanced.