摘要
通过分析传统方法研究AMT换档规律存在的问题和神经网络在不能获得精确数学模型的非线性系统中能达到最优控制的特性以及在线学习的能力等,提出基于广义回归神经网络(GRNN)进行AMT的换档规律的研究,并针对某4档轿车机械自动变速器,建立该车自动变速两个参数(车速、油门开度)神经网络控制模型,运用Matlab软件进行换档过程的仿真分析。研究结果表明:利用GRNN研究AMT的换档规律过程简单、适应性强等,能够正确有效地进行车辆档位判别。
Automated Mechanism Transmission （AMT） is focus of the automated transmissions＂ studying at home and abroad for closing structure ,high transmission＇s efficiency and the good produce inherit. The shifting rule of AMT is a complex ,nonlinearity mapping and classify question. It put forward the two parameters GRNN model of automatic transmission for car by artificial neural network control theory. The software MATLAB was employed to carry testing analysis and simulation. The simulation result shows that this method is accurate and feasible and is better than the result of BP.
出处
《机械设计与制造》
北大核心
2009年第5期72-74,共3页
Machinery Design & Manufacture
