基于人工神经网络的遥操作预测仿真

摘要

网络时延的存在给遥操作系统的感知和控制带来了许多问题。为减少网络时延对系统的影响，提出了一种基于人工神经网络的预测方法。预测仿真是克服基于网络的遥操作系统中网络传输时延的不确定性关键技术之一，结合这种预测方法，遥操作系统根据主端命令和从端反馈的数据，对远端的机器人状态进行预测，并在三维虚拟仿真场景中显示。实验结果表明，此预测仿真方法具有一定的可行性。%Internet time delay is the key inherent problem in teleoperation system. To reduce the uncertainty of time delay existing in internet teleoperation, a predictive simulation approach based on artificial neural network is proposed. According to master command and slave feed-back, teleoperation system predicts the state of the robot and displays in three-dimensional simulation system. Experimental results demonstrate the feasibility and validity of this approach.