论文部分内容阅读
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。
The neural network and fuzzy theory are comprehensively used. The characteristics of force feedback and active compliance control are analyzed in detail. A new strategy of “force / bit and loop control” is put forward for the first time. The corresponding neural network structure is established and the fuzzy rules based on knowledge The fuzzy coarse correction is implemented to eliminate the obvious interference signal; then the force / bit is effectively integrated by the characteristics of the neural network and directly input to the robot position servo system to implement the force / position and loop control. And the AdeptThree precision assembly robot carried out a very difficult experiment of scrubbing plane glass, effectively controlling the force within the ideal range of 8 ± 0.5N, and achieved satisfactory results.