基于残差决策的乒乓球发球机卡球故障自动检测方法

Automatic detection method of ball jamming fault of table tennis service machine based on residual decision

针对常规卡球故障检测方法获取的故障信号频域存在大量极值,导致最终检测所需的时间过长,针对该问题,设计一种基于残差决策的乒乓球发球机卡球故障自动检测方法。将获取的卡球故障信号分解为多个本征模式分量后,筛选故障信号本征极值,得到卡球故障信号对应的残差分量,在神经网络的支持下,利用残差分量构建一个检测优先的决策网络,计算得到故障检测决策参数,批量化处理决策参数后,建立一个故障构造函数,得到一个自动化检测过程。使用已知参数的发球机,并针对不同卡球位置设定故障数据集,准备两种常规检测方法与设计的自动检测方法进行实验,结果表明:设计的自动检测方法在检测过程中所需的时间最短。

In view of the large number of extreme values in the frequency domain of the fault signals obtained by the conventional fault detection methods, the final detection time is too long. To solve this problem, an automatic detection method for the sticking ball fault of table tennis machine based on residual decision is designed. Will get CARDS ball fault signal is decomposed into several intrinsic mode functions, screening of fault signal intrinsic extremum, get the residual components of the fault signal, the ball with the support of the neural network, using the residual component to build a testing priority decision network, decision-making parameters to calculate the fault detection, mass after the processing parameters, establish a fault constructor, get an automated test process. Two kinds of conventional detection methods and the designed automatic detection method were prepared by using the known parameters of the service machine, and the fault data set was set according to different card positions. The results show that the designed automatic detection method takes the shortest time in the detection process.