摘要
为了使机器人的行动能适应温室的复杂情况,提高机器人的运动控制精度,将模拟PID控制算法离散化,对编码反馈误差进行归一化处理,并将其输入3层的BP神经网络,研究了隐含层加权系数的计算方法,完成并验证了模糊BP神经网络PID控制系统的算法。结果证明,在同一指令时间内,模糊神经网络控制器能够很好地完成角度指令的零误差调节,与常规PID控制器相比较,模糊神经网络控制器的超调量显著减小。
In order to make the robot adapt to the complex situation of greenhouse, improve the precision of move controlling, analog PID was discrete, the feedback error of coding was normalized, and three-layer BP neural network was inputted in the research, the calculating method of hidden-layer addition coefficient was introduced, algorithm of BP neural network fuzzy PID controlling was completed and verdict. The result of test showed in the same instruction time, neural network fuzzy controller could achieve zero error measurement about angle instruction, and neural network fuzzy controller remarkably minish overtop value compared with the general PID controller.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2006年第2期87-90,共4页
Transactions of the Chinese Society for Agricultural Machinery
基金
"十五"国家科技攻关计划资助项目(项目编号:2004BA524B0302)
关键词
温室
机器人
驱动控制
算法
Greenhouse, Robot, Drive controlling, Algorithm
