摘要
综合运用了神经网络、模糊理论，详细分析了力反馈和主动柔顺控制的特点，首次提出了“力／位并环控制”的新策略，建立了相应的神经网络结构，利用基于知识的模糊规则实现模糊粗校正，消除明显的干扰信号；再利用神经网络特点将力／位有效综合，并直接输入机器人位置伺服系统，实施了力／位并环控制。并在ＡｄｅｐｔＴｈｒｅｅ精密装配机器人上进行了难度极大的擦洗平面玻璃的实验，有效地将力控制在８±０．５Ｎ的理想范围内，取得了满意的效果。
A novel strategy of force and position parallel control is firstly and systematically presented under the analyses of force feedback and active compliance control based on intelligent theories such as fuzzy theory and neural network. The related neural network structure is built and its algorithm is designed. Some experimental researches about wiping a plane glass, which is broken if the force exerted on its surface can′t be effectively controlled, have been implemented in an AdeptThree robot manipulator, and the contact force is controlled within 8±0.5 N. The results prove that the strategy of force and position parallel control is effective and feasible.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
1997年第3期251-256,共6页
Journal of Nanjing University of Aeronautics & Astronautics
