Modified Extremum Seeking Control and Its Application for Engine Power Train

(1)在Maple平台上建立了含发动机换档的单缸发动机的MLD模型。(2)将仿真神经网络学习律应用于MDL模型，但即使改变中间类和一种神经网络类型，也不能获得良好的控制性能。(3)对开关型(增益调度)的PID控制进行了仿真，得到了一般的控制效果，但很明显，调度点的选择和良好的PID参数整定是必要的。(4)对改进的极值寻优控制理论进行了理论检验(包括内模使用、时间积分误差的使用、当(4)与控制系统构成时，由于路面有变化，因此能够缩短制动距离。

(1) The MLD model of the single-cylinder engine including the engine shift is constructed on MAPLE.(2) Simulated NN-Learning law is applied for the MDL model, but it was not able to get good control performance even if changed the middle class and an NN type in various ways.(3) Simulated the switching type (gain scheduling) PID control, and a result is provided an ordinary performance, but it became clear that schedule point choice and good PID parameter tuning of the precision were necessary beforehand.(4) Performed the theoretical examination of the modified extreme seeking control theory (including the internal model use, the time integral calculus error use, and the feedback variable gain use).(5) When constituted (4) with control system, and there was the road surface change, depending was able to shorten braking distance as a result in (3).