In this paper, a remote-control system for a six-degree-of-freedom robotic arm that uses an indirect teaching method is proposed. In the indirect teaching method, an essential time delay occurs, which degrades the system performance. To overcome this time delay, which can be modeled using a Smith predictor, a model neural network (MNN) has been adopted. The Smith predictor is a model-based algorithm that is uncertain and prone to interference. In this study, the MNN has been utilized in an effective manner to model the system to support the Smith predictor algorithm. Using this time delay compensation, the outer loop proportional, integral, and derivative (PID) control gains are adjusted in an optimal manner through a PID neural network (PIDNN) to ensure that the robotic arm follows human commands precisely. By using the PIDNN proposed in this paper, the time required for indirect teaching application of the robot arm can be reduced.

本文提出了一种采用间接示教方法的六自由度机械臂遥控系统。在间接示教方法中，会出现必要的时间延迟，从而降低系统性能。为了克服这种可以使用史密斯预测器建模的时间延迟，采用了模型神经网络 (MNN)。Smith 预测器是一种基于模型的算法，具有不确定性且容易受到干扰。在本研究中，MNN 已被有效地用于对系统进行建模以支持 Smith 预测器算法。使用这种时间延迟补偿，外环比例、积分和微分 (PID) 控制增益通过 PID 神经网络 (PIDNN) 以最佳方式进行调整，以确保机械臂精确地遵循人类命令。