In many applications, it is very important to detect outliers during the analysis of normal data. Many existing methods preprocess the data to remove the outliers and then analyse the data accordingly when the data are contaminated by different unexpected outliers; however, it is difficult to use this method for applications where the data must be analysed online. In this article, we present an online monitoring approach for statistical process control (SPC) that is robust with respect to the presence of outliers. The Monte Carlo simulation results show that the proposed control chart is quite robust under the standard normally distributed data and, moreover, the control limit is not be affected by the number and sizes of the outliers. Furthermore, a real data example from a foetal heart rate (FHR) process is used to illustrate an application of our proposed procedure.

在许多应用中，在分析正常数据时检测异常值非常重要。许多现有的方法对数据进行预处理以去除异常值，然后在数据被不同的意外异常值污染时对数据进行相应的分析；但是，对于必须在线分析数据的应用程序，很难使用这种方法。在本文中，我们提出了一种用于统计过程控制 (SPC) 的在线监控方法，该方法对于异常值的存在是稳健的。Monte Carlo 仿真结果表明，所提出的控制图在标准正态分布数据下具有很强的鲁棒性，而且控制限不受异常值数量和大小的影响。此外，