Abstract: A intellisent vehicle faul diagnosis system ANN-based has been developed and related self-learn-ing algorithm of the syStem has been given Comparing with ES of vehicle fault diagnosis,the ANN-based dlagnosis system for Vehicle fault is better in Knowledge representation and inference procedureA actual ANN-based intelligent diagnosis system for vehicle motor faults has been designed,which in-cludes two hidden layers,The network has been trained for 500 times using 24 groupe of trainin8 vec-tor pairs,the results show that the success rate for fault diagnosis is more than 90%;

Key words: fault diagnosis, intelligent system, artifical neural network, expert system

摘要： 本文描述了基于人工神经网络的汽车故障智能诊断系统，提出了系统的自学习算法；根据汽车发动机故障的层次的特征，建立了系统的网络模型；在对比汽车故障诊断专家系统基础上，阐述了神经网络系统在知识表示与推理过程中的特点和优势。本文在微机上开发了包含两层隐含层的一个基于神经网络的汽车故障智能诊断系统原形，在２４组训练数据５００次循坏训练后，实验表示该系统诊断正确率为９０％。

关键词: 故障诊断, 智能系统, 人工神经网络, 专家系统