汽车AMT离合器起步控制系统的研究与测试

Research and Test the Control System of AMT Clutch in Starting Phase

针对汽车AMT自动变速器时滞非线性的复杂起步工况,根据熟练驾驶员经验提出了能够充分反应起步意图和实时状态的模糊控制策略,并采用四输入分层设计以加快搜索速度。用遗传算法对第二层模糊控制器的隶属函数进行优化,并在MATLAB/Simulink中建立了系统仿真模型,对优化前后的控制性能进行了对比,优化后的模糊控制器控制效果更好。将模型验证及代码SIL测试之后与手写代码集成,下载到硬件系统中并在AMT试验台架上进行测试,其控制性能与仿真结果一致。

To adapt to complex conditions in starting phase of AMT （Automated Mechanical Tranmission）,the fuzzy control strategy is proposed according to the skilled drivers＇ experience.The strategy can fully represent the driver＇s intention and real-time status.The controller has four inputs and two-layer structure to improve the search speed.Genetic algorithm is used to optimize the membership functions of the second layer.The simulation model is constructed in MATLAB/Simulink.The results of general fuzzy controller and GA-fuzzy controller are compared,while the GA-fuzzy controller gets better.Then the handwritten code is integrated with the automatically generated code after SIL,and the integrated code is downloaded to the hardware.GA-fuzzy controller is tested on AMT testbench,and the test result is same to the simulation.