基于动态修正贝叶斯模型的自动测试系统量值稳定性监控方法

在两次计量确认间隔期内，借鉴测量过程控制思想，提出一种自动测试系统量值稳定性监控方法，用以保障其性能指标。在核查数据为小样本情况下，针对传统控制图方法在测量过程控制中容易导致“误发警报”的问题，采用动态修正贝叶斯模型进行核查数据概率分布参数的优良估计，从而改进控制图的控制参数。仿真分析表明，改进后的控制参数明显优化，降低了“误发警报”错误的概率，监控效果明显改善。

A Method of Monitoring the Quantity Value Stability for Automatic Test System Based on Dynamic Revised Bayesian Model

Using measurement process control principle for reference, a method of monitoring the quantity value stability was proposed for automatic test system to assure its precision in the interval of two metrological confirmations. When check data had the characters of small sample, the traditional control chart increased the probability of error alarm . Therefore, a dynamic revised Bayesian model was adopted to estimate the parameters of the distribution. Then the control parameters of traditional control chart were improved. The experiments show that the improved control parameters of control chart are optimized distinctly and the probability of “error alarm” is decreased.