基于广域网的遥操作工程机器人神经网络控制  

Neural Network Control for Tele-operated Construction Robot Based on WAN

作　　者：秦辉明 李笑[1] 

机构地区：[1]广东工业大学机电工程学院,广东广州510006

出　　处：《机床与液压》2014年第3期5-8,共4页Machine Tool & Hydraulics

基　　金：广东省"211工程"重点学科建设项目(412110901)

摘　　要：针对在广域网环境下因较大网络时延造成遥操作工程机器人系统操作性能下降甚至系统不稳定的问题,提出一种神经网络控制方法。借助于广域网模拟器,在局域网环境下进行了不同网络时延下的位置阶跃响应和稳态跟踪实验研究。结果表明:与常规PID控制方法相比,该方法具有更好的动态和稳态特性,对网络时延具有较好的适应性和鲁棒性。研究结果对提高广域网环境下的遥操作工程机器人系统的操作性能具有指导意义。The large network time-delay in the WAN environment degrades the performance of the tele-operated construction ro-bot system,and even make the system unstable. To solve this problem,a neural network control method was proposed. With the help of WAN emulator,the experimental researches of position step response and steady state tracking under variable network time-delay were conducted in the LAN environment. The experimental results show that the proposed method has better dynamic and steady state characteristics and has better adaptability and robustness to the network time-delay,compared with the conventional PID control meth-od. The research result has guiding significance for improving the performance of tele-operated construction robot system in the WAN environment.

关 键 词：遥操作 工程机器人 广域网 神经网络 PID控制 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置] TH137[自动化与计算机技术—控制科学与工程]