为抵消网络可变时延对遥操作系统的影响,基于一种新的改进型Smith预估器,构建了预估控制下的网络实时遥操作移动机器人系统.该系统根据主、从端传感器交换的信息,通过动态模型管理器及其算法,保证了主、从端模型的一致性.文中根据端到端数据包多个到达的特性,提出了新的时延缓冲器管理算法,从而更好地将可变时延转化为常时延.为了增强系统控制的实时性,系统中还引入了模型的虚拟显示.长距离的网络遥操作实验验证了文中系统和控制策略的实用性及有效性.
In order to eliminate the impact of variable network delay on teleoperated systems, this paper presents an Internet-based real-time teleoperated mobile robot system by utilizing the predictive control scheme and introducing an improved Smith predictor. Based on the sensor information exchange between the master and the slave, the system guarantees the master and slave models synchronism through a dynamic model manager and its algorithm. Moreover, according to the many arrivals of end-to-end datagram, a new regulating algorithm for delay buffer is pro- posed to transfer a variable delay into the constant one. Virtual model display is also introduced to enhance the realtime efficiency of the system. The usefulness and effectiveness of the developed method and system are finally validated by the teleoperation experiments through long-distance Internet.