Abstract Colonoscopy is a common procedure to perform advanced therapies such as Endoscopic Submucosal Dissection (ESD), which allows for greater diagnostic specificity and sensitivity compared to other types of examination. Nevertheless, since the colonoscope can cause patient discomfort or pain due to improper manipulation, it is quite challenging for endoscopists who need to develop the necessary skills to accurately perform the procedure and minimize this discomfort. To overcome these sorts of limitations inherent in conventional colonoscopies, various studies regarding robotic and automated systems have been made. In this paper, based on the mechanics of paddling locomotion, a fully self-propelled robotic colonoscope is proposed, then integrated with assistance modules. The feeding and the steering module aim to assist the navigation of the human colon. By building on operations already familiar to endoscopists (i.e., tip deflection, push forward and pull back, hooking, etc.), the device's automatic sequences are investigated to ensure high safety and maneuverability. In addition, a Multimodal Robotic Colonoscope Interface (MRCI) is integrated with each modular mechanism, which allows endoscopists to monitor kinetic/kinematic data and implement manual control in case of unexpected emergency (i.e., deadlocked in colon, paradoxical movement, etc.). With the integrated interface, a pilot study using a porcine colon was conducted. Through an In-vitro test along a straight path, the velocity of the proposed RC was identified as 16.9 mm/s at a paddling frequency of 2 Hz with feeding force of 10 N. These results are 1.7 times faster (9.6 mm/s) than tests that only used paddling locomotion without any feeding force and steered angle. Furthermore, at a curved path with a radius of 60 mm, the RC velocity was measured at 21.5 mm/s, and experienced a radial elongation ratio of 20%. Similarly, at an inclined path of 30°, the velocity of the robot increased 25% (8.93 mm/s) compared to paddling locomotion alone (7.14 mm/s). The final results showed that the integrated system had superior results compared to previous studies based on self-propelled locomotion alone.

中文翻译：

---

一种新型机器人结肠镜检查系统，将喂食和转向机构与自推进桨运动相结合：一项初步研究

摘要 结肠镜检查是执行内窥镜黏膜下剥离术 (ESD) 等高级治疗的常见程序，与其他类型的检查相比，它具有更高的诊断特异性和敏感性。然而，由于结肠镜会因操作不当而导致患者不适或疼痛，因此内窥镜医师需要掌握必要的技能来准确地执行手术并将这种不适感降至最低，这对内窥镜医师来说是非常具有挑战性的。为了克服传统结肠镜检查中固有的这些限制，已经进行了有关机器人和自动化系统的各种研究。在本文中，基于划桨运动的力学，提出了一种完全自走式机器人结肠镜，然后与辅助模块集成。喂食和转向模块旨在辅助人类结肠的导航。通过建立内窥镜医师已经熟悉的操作（即尖端偏转、向前和向后推、钩等），研究设备的自动序列以确保高安全性和可操作性。此外，多模式机器人结肠镜接口 (MRCI) 与每个模块化机构集成在一起，这使内窥镜医师能够监测动力学/运动学数据并在意外紧急情况（即结肠僵局、反常运动等）的情况下实施手动控制。借助集成界面，进行了使用猪结肠的试点研究。通过沿直线路径的体外测试，在 2 Hz 的桨频和 10 N 的进给力下，所提出的 RC 的速度被确定为 16.9 mm/s。这些结果比仅使用划桨运动而没有任何进给力和转向角的测试快 1.7 倍（9.6 毫米/秒）。此外，在半径为 60 mm 的弯曲路径上，RC 速度测量为 21.5 mm/s，径向伸长率为 20%。同样，在 30° 的倾斜路径上，机器人的速度与单独的划桨运动 (7.14 mm/s) 相比增加了 25% (8.93 mm/s)。最终结果表明，与先前仅基于自走运动的研究相比，集成系统具有更好的结果。在 30° 的倾斜路径下，与单独的划桨运动 (7.14 mm/s) 相比，机器人的速度增加了 25% (8.93 mm/s)。最终结果表明，与先前仅基于自走运动的研究相比，集成系统具有更好的结果。在 30° 的倾斜路径下，与单独的划桨运动 (7.14 mm/s) 相比，机器人的速度增加了 25% (8.93 mm/s)。最终结果表明，与先前仅基于自走运动的研究相比，集成系统具有更好的结果。