发表机构
Imperial College London（帝国理工学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究针对机器人辅助微创手术中禁区主动约束（FRAC）的动态组织变形等挑战，提出能量耗散型FRAC策略，经体外实验验证其鲁棒性、有效性及时间效率优于对比方法。
AI 中文摘要
在机器人辅助手术中，禁区主动约束（Forbidden Region Active Constraints, FRAC）是一种通过向外科医生生成各向同性触觉引导来维持任务安全性的控制策略。然而，FRAC在临床环境中用于远程手术前需克服多项挑战，包括允许动态组织变形、维持能量被动性及实现速度等。本研究提出一种能量耗散型FRAC策略的流水线设计，利用深度感知相机适配呼吸运动引起的动态组织变形。该FRAC策略采用精细网格表示，案例研究中总多边形数为122806个，运行频率达43.48Hz。我们设计了由“虚拟”外科医生进行的体外轨迹跟踪实验以辅助对该方法的定量评估，包括其维持任务安全性的有效性，所有实验均成功保持预定义安全距离，验证了这一点。我们还开展了对比研究，以调查该方法相对于其他依赖简单几何主动约束（AC）表示的FRAC方法的鲁棒性和时间效率，结果表明，本方法总体提供更鲁棒、有效的引导，同时保持相当甚至更低的时间成本。
英文摘要
In robot-assisted surgery, Forbidden Region Active Constraints (FRAC) represent a control strategy that helps maintain task safety by generating anisotropic haptic guidance to surgeons. However, several challenges need to be overcome before FRAC can benefit teleoperative surgery in a clinical setting. These challenges include the ability to allow for dynamic tissue deformation, maintain energetic passivity, and speed of implementation, among others. In this study, we propose the pipeline design for an energy dissipative FRAC strategy, which accommodates the dynamic tissue deformation caused by respiratory movements, by utilizing a depth sensing camera. The proposed FRAC strategy adopts a fine mesh representation, with a total number of 122,806 polygons in the case study presented, while running at 43.48Hz. We designed in vitro trajectory tracking experiments conducted by a "virtual" surgeon to aid quantitative assessment of the method, including its effectiveness in maintaining task safety, which was confirmed by successfully maintaining a pre-defined safety distance across all trials. We also conducted comparative studies to investigate the robustness and time-efficiency of our method against other FRAC methods that rely on simple geometry AC representations. We demonstrate that our method provides a more robust and effective guidance overall, while maintaining comparable, if not lower, time costs.
Comments8 pages, 12 figures
Journal refVolume: 10, Issue: 3, March 2025, Page(s): 2950 - 2957
