论文链接：https://ieeexplore.ieee.org/document/11072297
视频链接：https://ieeexplore.ieee.org/document/11072297/media#media
传统的CT影像设备通常为O型结构或单机械臂加C型臂的配置形式，其体积重量大、扫描轨迹灵活性差。为此，我们提出了一种新颖的双机械臂CT扫描系统，两个机械臂末端分别安装了X射线发射球管和X射线平板探测器，其具备扫描空间开放，扫描轨迹灵活等诸多优势。另一方面，工业CT和医用CT均期望以最短时间完成扫描以提升扫描效率或减少病人的辐射剂量，为此，本文提出了一种考虑任务约束和同步运动约束的双机械臂CT扫描系统最短时间同步轨迹规划算法。该算法能够在满足给定的双臂任务约束和同步运动约束的前提下规划出双臂时间最优轨迹。实际的CT成像结果表明，基于对比方法所获得的CT图像具有严重的运动伪影，而基于本文方法获得的CT图像消除了运动伪影且可清晰显示最小半径为1毫米的模拟血管，从而证明了本文所提方法的优越性。
H. Wang, M. Yuan*, J. Huang, X. Zhang and F. Liu*, "Minimum-Time Trajectory Planning of a Dual-Manipulator CT System With Synchronization and Task Constraints," in IEEE Robotics and Automation Letters, vol. 10, no. 8, pp. 8475-8482, Aug. 2025, doi: 10.1109/LRA.2025.3586521.
Abstract
In industrial and clinical applications, CT scanning is expected to be as fast as possible to improve the scanning efficiency in industrial CT or to reduce the radiation dose to patients in clinical CT. However, the kinematic constraints of the dual manipulators pose limits on shortening the scanning time. In addition, the X-ray tube and detector mounted on the end-effectors of the dual manipulators must be precisely aligned during the scanning motion, which imposes the synchronization motion constraint on the dual manipulators. In this letter, considering the kinematic constraints (velocity and acceleration constraints) in the task space of dual manipulators and the synchronization motion constraint, we propose a minimum-time trajectory planning algorithm for a dual robotic manipulator CT system. Specifically, the CT scanning paths of both the X-ray tube and detector are formulated as a parametric form. Then, a novel unified bound estimator is proposed to convert the task constraint and synchronization constraint into the parametric constraints. Finally, the minimum-time trajectory planning issue of the dual manipulators is solved in real time by a nonlinear filter in the parameter space. Comparative experiments are conducted on two UR20 manipulators. The experimental results confirm that the planned trajectories by the proposed approach are time-optimal and satisfy all the task and synchronization motion constraints as well.
