基于K—L变换的汽车电控系统技术状态特征选择与提取的研究

随着汽车电子化程度的日益提高，汽车电控系统的故障诊断方法和理论受到了越来越广泛的关注。基于模式识别技术的汽车电控系统故障诊断方法具有非常广阔的应用前景，但是对具体的事件进行模式采集时，有可能造成样本在模式空间中的维数很大．由此带来数据处理的困难。若可通过K—L变换方法对汽车技术状态的特征参数进行选择和提取，并对选择前后的特征分别用神经网络进行模式识别，不仅能够提高识别的速度，而且能够提高识别的精度。

Technical State Feature Selection and Extraction of Electrically Controlled Automobile Based on K-L Transformation

With the improvement of automobile electrical degree , more and more people begin to pay attention to the fault diagnosis methods and theories of electrically controlled system. The method based on pattern recognition has bright application prospects. Collecting pattern data may bring about large dimensions and difficult data process. K-L transformation is used to select and extract automo- bile technical state parameter and neural network is applied to recognize the features before extraction and after extraction. The result shows that the method can not only heighten the recognition velocity, but also promote recognition precision.