中文摘要：

在汽油机瞬态空燃比反馈控制过程中，氧传感器存在传输时滞，不能快速反馈汽油机瞬态空燃比真实值，无法满足瞬态空燃比反馈控制的实时性要求。文章提出了汽油机瞬态空燃比的混沌时序LS-SVM（最小二乘支持向量机）预测模型，采用相空间再构技术对原始数据进行重构，达到恢复汽油机瞬态空燃比时间序列的多维空间非线性特性目的，最后利用LS-SVM进行训练及预测，得到空燃比预测结果。仿真结果表明，与Elman网络及前馈BP网络相比，混沌时序LS-SVM预测模型具有更强的非线性预测能力，能够有效地提高瞬态空燃比的预测精度，为瞬态空燃比反馈控制的成功实行提供了有力的依据。

英文摘要：

In the process of feedback control of gasoline engine transient air-fuel ratio, the oxygen sensor has transmission delay and can not feed back the true value of gasoline engine transient air-fuel ratio quickly, thus failing in real-time control of transient air-fuel ratio. In this paper, the chaotic time series least squares-support vector machine（LS-SVM） prediction model of the gasoline engine transient air-fuel ratio is proposed. First, the original data are reconstructed by using phase-space reconstruction technique so as to recover the multidimensional nonlinear characteristics of time sequence of gasoline engine transient air-fuel ratio. Then LS-SVM is applied to training and identifying the reconstructed data. Finally, the air-fuel ratio identification results are obtained. The simulation results show that compared with the Elman neural network and feedforward BP neural network prediction models, the chaotic time series LS-SVM prediction model has stronger nonlinear prediction capability, and it can improve the prediction precision of transient air-fuel ratio effectively. This study can provide a basis for precise feedback control of transient air-fuel ratio.