力控制反馈信号的神经网络法校正

Rectification of Force Feedback on Active Compliance Control by Neural Network

对机器人的力反馈、主动柔顺控制和双向联想记忆的神经网络进行了综合研究，针对机器人腕力传感器输出信号的部分失真，提出了机器人擦洗平面玻璃力反馈信号神经校正的方法，并建立了相应的神经网络结构，在ＡｄｅｐｔＴｈｒｅｅ精密装配机器人上进行了试验验证，取得了满意的效果。

The theories of force feedback,active compliance control and bidirection associate memory (BAM) are synthetically studied in detail. To deal with the problem of force feedback on SAFMS, a neural method is developed. By this method, the BAM structure of the AdeptThree manipulator wiping plane glass is built with good results.