中文摘要：

提高电控汽油机空燃比控制精度是改善发动机燃油经济性、动力性和降低尾气污染的关键环节．针对稀薄燃烧汽油机的工作原理，提出了一个稀燃汽油机空燃比滑模．神经网络控制方案，并对方案中的各环节进行了详细描述．采用自行开发的发动机电控系统，在一台稀燃发动机上进行了实验，并对实验结果进行了分析．实验结果表明，采用滑模-神经网络方案对稀薄燃烧发动机空燃比进行控制，不仅可以提高准稳态时发动机的空燃比控制精度，而且可以降低过渡过程的空燃比超调．节气门急速变化时的空燃比超调最大为1个空燃比单位，最小为0．2个空燃比单位，大大优于车用电控系统的控制结果．

英文摘要：

To enhance the controlling accuracy of engine air-fuel ratio （ AFR）, improve the fuel economy and powering ability and reduce the pollution of tail gas, a sliding mode-neural network control scheme for the air-fuel ratio of the lean burn engine is presented for the characteristics of the lean burn engine. With the engine control system developed by ourselves, the lean burn engine air-fuel ratio control experiments are carried out on a lean burn gasoline engine. The results show that applying sliding mode-neural network scheme to controling the air-fuel ratio of the lean burn engine, both the steady-state and transient air-fuel ratio control accuracy can be enhanced. The maximal overshoot of air-fuel ratio is 1 AFR unit and the minimal overshoot reduces to O. 2 AFR unit when the throttle body moves rapidly, completely better than the vehicle used electronic control unit （ECU）.