基于半参数回归模型的微陀螺仪温度补偿技术

设计了一种基于半参数回归模型的微陀螺仪温度补偿方法.针对微陀螺仪启动时受温度变化影响明显、角速度输出误差较大的问题,提出对于温度的补偿不仅要考虑趋势项,而且要估计模型误差.建立了温度补偿的半参数回归模型,采用补偿最小二乘进行求解.实验结果表明,对于某型微陀螺仪采用温度补偿的半参数回归模型是可行且有效的,相对于常规的多项式拟合方法,补偿后的信号噪声降低了1个数量级,优于0.005 (°)/s,从而有效缩短了陀螺仪启动稳定时间.

Study on Temperature Compensation for MEMS Gyro Based on Semiparametric Regression Models

A temperature compensation method for MEMS gyro based on semi-parametric regression models has been developed.Angular rate of gyro fluctuates dramatically with the rising of gyro temperature when gyro starts up.In this paper,it is suggested that temperature compensation model should consider not only the trend term,but also the model error term.According to the analyzed results,a semi-parametric regression model of temperature compensation is designed for a certain kind of MEMS gyro.The solution of the model is studied using the method of penalized lest square.Experimental results show that the temperature compensation method is feasible and effective.The perfect performance with accuracy of better than 0.005 (°)/s can be achieved and the errors of the presented method are depressed significantly for an order of magnitude than the traditional method of polynomial fit.It is concluded that start-up time of gyro is shortenen when adopting the proposed temperature compensation model.