CAJ | 학술논문

针对机械臂运动轨迹控制中存在的跟踪精度不高的问题，采用了一种基于EC-RBF神经网络的模型参考自适应控制方案对机械臂进行模型辨识与轨迹跟踪控制。该方案采用了两个RBF神经网络，运用EC-RBF学习算法，采用离线与在线相结合的方法来训练神经网络，一个用来实现对机械臂进行模型辨识，一个用来实现对机械臂轨迹跟踪控制。对二自由度机械臂进行仿真，结果表明，使用该控制方案对机械臂进行轨迹跟踪控制具有较高的控制精度，且因采用EC-RBF学习算法使网络具有更快的训练速度，从而使得控制过程较迅速。
침대궤계비운동궤적공제중존재적근종정도불고적문제，채용료일충기우EC-RBF신경망락적모형삼고자괄응공제방안대궤계비진행모형변식여궤적근종공제。해방안채용료량개RBF신경망락，운용EC-RBF학습산법，채용리선여재선상결합적방법래훈련신경망락，일개용래실현대궤계비진행모형변식，일개용래실현대궤계비궤적근종공제。대이자유도궤계비진행방진，결과표명，사용해공제방안대궤계비진행궤적근종공제구유교고적공제정도，차인채용EC-RBF학습산법사망락구유경쾌적훈련속도，종이사득공제과정교신속。
According to the problem that the tracking accuracy is not high enough in trajectory tracking control of robot manipulators, a model reference adaptive control scheme based on EC-RBF neural networks is adopted to achieve robot manipulator model identification and trajectory tracking control. This control scheme contains two RBF neural networks which are trained offline and online, using EC-RBF learning algorithm. The one is used to identify the robot manipulator’s model, and the other one is used to achieve its trajectory tracking control. Simulation result of 2-degree-of-freedom robot manipulator demonstrates that using this method for robot manipulator trajectory tracking control has high control accuracy, and the networks which gain high training speed because of the EC-RBF learning algorithm make the control process faster.