基于小波变换的学前儿童注意力测试仪故障检测方法

Fault detection method of preschool children’s attention tester based on wavelet transform

针对传统学前儿童注意力测试仪故障检测准确率低,导致注意力测试效果不佳的问题,提出基于小波包变换的学前儿童注意力测试仪故障检测方法。首先利用小波包分解方法对采集信号进行消噪;然后通过能量谱对测试仪初始故障特征向量进行提取;之后采用主成分分析方法对提取特征进行降维处理,从而得到新的故障特征向量;最后采用L-M算法改进的BP神经网络对故障特征进行识别和分类。实验结果表明,相较于其他信号消噪预处理方法,提出的最优预测变量阈值法的消噪效果最好,可为后续故障特征提取和识别检测提供有效数据。仿真测试证明,在检测仪触发角分别为0度、30度和60度时,改进BP神经网络的故障诊断精度分别取值为90%、92.72%和88.18%,总诊断率为90.3%。由此可知,提出的故障检测方法可取得较高的检测准确率,可提升注意力测试仪检测效果,故障诊断效果较好,可在儿童注意力测试仪中进行广泛应用。

In view of the problem of low fault detection accuracy of traditional preschool attention tester, the fault detection method of preschool attention tester based on wavelet packet transformation is proposed. The wavelet packet decomposition method is used to eliminate the collected signal noise;then the initial fault feature vector is extracted through the energy spectrum, then the principal component analysis method is used to reduce the new fault feature vector, and finally the BP neural network improved by L-M algorithm is used to identify and classify the fault features. The experimental results show that, compared with other signal noise cancellation preprocessing methods, the proposed optimal predictor variable threshold method has the best noise cancellation effect, which can provide effective data for subsequent fault feature extraction and identification and detection. Simulation tests proved that the fault diagnosis accuracy was 90%, and 88.18% were 90%, 30% and 88%, and the total diagnosis rate was 90.3%. Therefore, we can see that the proposed fault detection method can achieve high detection accuracy, which can improve the detection effect of the attention tester, and the fault diagnosis effect is good, and can be widely used in the children’s attention tester.