车辆起步过程神经网络PID控制研究

针对车辆起步工况的多样性, 笔者设计了神经网络PID(Proportion Integration Differentiation)控制器, 通过对离合器的转速跟踪控制, 在保证发动机不熄火的前提下, 实现了AMT(Automatic Manual Transmission)车辆快速平稳起步的控制目标。为了验证控制器的有效性,建立了中型卡车的传动系模型, 其中包括发动机、 离合器、 变速箱、 主减速器和车轮等, 并且根据实际起步情况, 设置了正常起步、 急起步、 载重起步和坡路起步4种工况, 进行了AMESim-Simulink联合仿真实验。通过与传统PID控制器的控制效果相比较, 分析得出,神经网络PID控制器能在变工况下很好地完成AMT车辆起步控制, 具有良好的自适应性。

Research on Launch Control Strategy for AMT Vehicles Based on Neural Network PID

According to the diversity of launch conditions a neural networkPID (Proportion Integration Differentiation) controller is designed through the clutch speed tracking control. Keeping the engine running, the controller makes the launch process of the AMT (Automatic Manual Transmission) vehicle smooth and fast. In order to verify the neural network PID controller's effectiveness, a powertrain model of a medium-sized truck about AMT vehicles is established. The model includes the engine, the clutch, the transmission, the main reducing gear and wheels. And simulations are carried out in various working conditions according to the actual situation. Comparing the performance to the traditional PIDcontroller's and analyzing the results of simulations, the neural network PID controller provides the good performance during AMT vehicles launch progress and has good adaptability in various working conditions.