基于相关向量机的加速度计参数变化规律研究

Research on Parameters of Accelerometer Based on Relevance Vector Machine

针对温度、振动等多种影响因素影响石英挠性加速度计内部参数变化,使石英挠性加速度计参数随着时间发生漂移,从而难以准确描述其变化规律的问题,基于相关向量机具有稀疏的模型结构,适合进行不确定系统建模的优点,提出了一种采用相关向量机进行石英挠性加速度计参数规律研究的方法。为了验证方法的有效性,针对2.5年自然贮存条件的石英挠性加速度计的参数标定值进行了时间序列建模,计算结果表明相关向量机具有很好的效果。

Temperature and vibration and other environment factors have important effects on the inner parameters of quartz flexure accelerometer. The accelerometer parameters drift is occurred with time. The accelerometer parameters are difficult to be described exactly. Relevance vector machine which has better performance in sparsity,can be applied to modeling of uncertain systems. An accelerometer parameters prediction method based on the grey least squares support vector machine is proposed. The 2. 5 years natural environmental scale value of accelerometer parameters is modeled. The experimental results show that relevance vector machine has well performance.