摘要
遥操作机器人系统传输通道中存在通讯时延 ,造成系统不稳定和操作性能降低等问题 .为了消除或减少时延的影响 ,利用前向神经网络建立环境模型 ,通过神经网络模型预测从机械手受力 ,并用时间前向观测器预测从机械手状态 ;然后用力、位置和速度反馈对系统进行设计 ,使得系统在环境模型未知的条件下稳定 ,还能获得良好的透明性 ;最后 ,将预测方法与无源控制方法进行切换 ,保证系统在环境剧变时的稳定性 .
The communication time delay in communication channels of a teleoperation robot system conduces to instability and bad performance of the system. In order to eliminate or reduce the effect of time delays, back propagation neural network is used to build the model of environment. Force and states of the slave are predicted through the neural network model and time forward observer. Moreover, force, position and velocity feedback are proposed, control parameters are designed and controllers are switched between predictive control and passive control to ensure the stability and performance of the system. The experimental results prove the effectiveness of this method.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2003年第2期177-181,共5页
Journal of Southeast University：Natural Science Edition
基金
国家 8 63高技术资助项目 ( 863 2 0 0 1AA42 3 2 40 )
