光纤陀螺温度误差自适应神经模糊补偿方法  

Temperature Error Compensation Method Based on Adaptive Neuro Fuzzy Inference for Fiber-optic Gyro

作　　者：冯卡力 李安[1] 覃方君[1] 李峰[1] 

机构地区：[1]海军工程大学电气工程学院,湖北武汉430033

出　　处：《兵工学报》2016年第4期641-647,共7页Acta Armamentarii

基　　金：国防预先研究项目(4010103010101)

摘　　要：依据Shupe热致非互易性理论,研究光纤陀螺温度漂移产生机理,设计了-15℃～50℃温度范围内实验。将自适应神经模糊推理系统和分段建模思想相结合,以实测数据构建训练样本,建立光纤陀螺温度误差自适应神经模糊补偿模型。该模型解决了传统建模方法中模型适配性较差的问题,并有助于缩短陀螺进入稳态的时间。数据计算表明,用所建立的模型进行误差补偿后,温度漂移标准差降低了75.55%,补偿效果明显。According to Shupe thermally induced non-reciprocity noise theory,the mechanism of the temperature drift of fiber-optic gyro is researched,and a temperature experiment with range of- 15 ～ 50 ℃is designed. A temperature error compensation model based on adaptive neuro fuzzy inference for fiberoptic gyro is established by using the adaptive neuro fuzzy inference system and the piecewise modeling method and constructing the training sample from the measured data. The proposed model is used to solve the problem of poor model matching,and can help to shorten the time of the gyro entering into the steady state. The calculated results show that the standard deviation of the temperature drift is reduced by75. 55% after compensation with the proposed method.

关 键 词：控制科学与技术 光纤陀螺 温度漂移 自适应神经模糊推理 误差补偿 分段建模 

分 类 号：TG156[金属学及工艺—热处理]