基于动作追踪的遥操作控制及具身数据采集架构设计OA

Traditional teleoperation for space robots suffers from complex interaction and low op-eration efficiency.Distributed teleoperation systems also exhibit loose coupling between control and data-acquisition links,together with fragmented safety mechanisms.To obtain high-quality multimodal demonstration data for embodied-intelligence research of space robots,a motion tracking-based architecture for teleoperation control and embodied-data acquisition is proposed.Physical layering and dual-end communication-process isolation are adopted in this architecture.Three layers are defined:the motion-tracking and human-robot interaction layer,integrated con-trol and multimodal data-acquisition layer,and robot execution layer.Key functional modules are integrated,including spatial geometric mapping,constrained inverse kinematics solving,multi-level safety constraints,and synchronous multimodal-data acquisition.Relying on complete safe-ty-constraint mechanisms,the system realizes mapping from human natural hand motions to ro-bot control commands.It implements teleoperation control as well as high-quality synchronous embodied-data acquisition.A physical verification platform is built with a head-mounted motion-tracking terminal and a 7-degree-of-freedom manipulator.Test results verify that the system meets on-orbit control demands of space robots in tracking accuracy,dynamic stability and end-to-end real-time performance.The architecture enhances teleoperation dexterity and operation ef-ficiency for complex on-orbit tasks.It supplies high-quality expert demonstration data for embod-ied-intelligence model training,and provides technical references for engineering realization of space-robot intelligent control systems.