预估控制下的实时网络遥操作移动机器人

构建了能使操作者通过Internet远程实时控制的移动机器人系统．为了补偿网络时延和抵消其对遥操作系统的影响，基于我们以前提出的改进型Smith预估器原理，采用了预估控制策略．为了保证系统稳定性和透明性，基于主从端的传感器信息交换，设计了一个动态模型管理器，其中模型和力反馈误差调节通过模糊控制实现．除了力反馈外，为了增强遥操作的实时性，引入了预估的虚拟显示．为了精确地预测网络时延，提出了一个新颖的时钟同步算法．为了降低时延抖动，结合我们提出的两个算法，实现了数据缓冲策略．最后，通过长距离的网络遥操作实验验证了系统和控制策略的实用性和有效性．

Real-time Mobile Telerobot over IP Networks Based on Predictive Control

This paper presents a system that enables the remote operator to control a mobile robot with real-time control style via Internet. To compensate network delay and cancel its impact on teleoperation system, we adopt a predictive control scheme based on a modified Smith predictor we proposed previously. To guarantee system stability and transparency, a dynamic model manager isdesigned based on sensor information exchange between the master and slave sides, where model and force feedback errors are regulated with a fuzzy controller. In addition to force feedback, a virtual predictive display is introduced to enhance the real-time efficiency of teleoperation. To predict time delay precisely, a novel timer synchronization algorithm is proposed. To decrease delay jittering, a data buffer scheme is achieved with the two proposed algorithms. The usefulness and effectiveness of the developed method and system are validated by teleoperation experiments through Internet at a long distance.