对角神经网络用于环形激光陀螺消噪建模

Ring laser gyro noise elimination and modeling by diagonal recurrent neural networks

对角神经网络(DRNN)为非全反馈式动态神经网络。应用DRNN对处于静止和运动状态下的环形激光陀螺(RLG)进行了消噪建模,并应用A llan方差方法对消噪后的结果进行了对比分析。结果表明:使用DRNN对RLG进行消噪建模是可行的。同时,将DRNN与反向传播神经网络的消噪结果进行了比较,得到动态网络的消噪能力要优于静态网络的结论。所用方法对研究RLG的误差补偿及快速启动是有实际意义的。

Diagonal recurrent neural network （DRNN） is a non-unity feedback network. DRNN is used in ring laser gyro（RLG） noise elimination and modeling when it is static and dynamic. The method of Allan variance is used to analyze the gyro data before and after noise elimination. The results show that DRNN can be used to eliminate the noise of RLG, DRNN is compared with reverse network, and the dynamic network is prior to static network. The method is meaningful to the research of RLG error compensation and fast start.