摘要
针对工程车辆动力换挡系统多变量和非线性的特点,提出一种基于模糊神经网络的智能协调控制方案。为克服 传统的BP网络的不足,采用自适应变步长算法(ABPM)来训练前馈人工神经网络,并用TY165EH推土机换挡试验中获 得的数据进行了仿真实验,结果表明:该智能换挡系统克服了一般BP神经网络学习速率慢和陷入局部最优的缺点,可根 据推土机操作工况环境实现正确的换挡。
In order to improve the intelligence of engineering vehicle shift decision,a kind of adaptive Fuzzy Neuro Networks System (ABPM) was proposed and test simulation was developed based on the characteristics of multy-varibles and nonlinearity of engineering vehicle shift system and data obtained in shift experiment on TY165EH bulldozer transmission system.The simulation results show that the shift decision system can realize correct gear-box shift decision according to operation conditions.The controller overcome the shortcoming of Back-Propagation Neuro Network,the requirements of efficiency and global-optimization can be met at the same time.
出处
《机床与液压》
北大核心
2005年第3期161-163,共3页
Machine Tool & Hydraulics
关键词
工程车辆
换挡
模糊神经网络
Engineering vehicle
Shift
Fuzzy neuro networks
