Abstract: Automotive diagnostic equipment be able to detect faults automatically and use artificial intelligence technology. Machine learning enables the computer to learn useful knowledge from external input data to adjust the algorithm structure to improve the reliability of fault diagnosis. The improved extreme point symmetric mode decomposition method is used to extract the initial intrinsic mode function, then import it into multi-layer perceptron network for training to establish a failure mode classification system in this paper.

Key words: Bearing fault diagnosis; MLP; Machine learning; Fault mode classification

摘要： 汽车诊断设备重要特征是应用自动化的综合诊断技术提高对复杂故障的诊断和预测能力，使汽车检测与故 障诊断技术向人工智能方向发展。机器学习可以使计算机模仿生物机体从外部输入的信息数据中学习有用知识用于 调整优化自身算法结构提升故障诊断的可靠性。文章应用改进极点对称模态分解方法提取故障振动信号的初始模态 分量导入多层感知器（MLP）网络训练，建立故障模式识别系统。

关键词: 轴承故障诊断；多层感知器网络；机器学习；故障模式识别