SIA OpenIR  > 机器人学研究室
消化内镜手术机器人关键技术研究
Alternative TitleThe Research on the Key Technologies of Surgical Robot for Digestive Endoscopy
江维
Department机器人学研究室
Thesis Advisor王志东
Keyword消化内镜手术 控制策略 背隙建模 情景感知 自主介入
Pages116页
Degree Discipline机械电子工程
Degree Name博士
2021-05-21
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract经内镜逆行性胰胆管造影(Endooscopic Retrograde CholangioPancreatography)手术由于所作用的胰胆管区域在消化道深部,且涉及肝脏、胰腺等重要脏器,因而难度大、风险高,被誉为消化内镜手术皇冠上的明珠。目前临床中ERCP手术仍存在一些问题导致其对医生的技能和经验依赖程度较高,只能在一些三甲医院才能开展,主要包括手术器械不灵巧导致介入难度大、多人协作操控内镜和器械效率低以及辐射危害医护人员生命健康等问题。本论文在国家重点研发计划的资助下,针对消化内镜手术机器人构型设计、控制策略以及自主介入方法等关键技术进行深入研究。具体内容如下:(1) 分析ERCP介入任务需求,在先前研制的内镜操控机器人基础上,开展结构优化设计,模仿临床医生双手操控形式来配置内镜操控机器人的自由度。剖析经内镜逆行介入胆胰管的操作任务对手术臂机械结构以及运动特性要求,研制连续体型镜下手术臂,该手术臂能够沿消化内镜的器械腔道抵达十二指肠部位,完成精细定向插管任务,并逆行介入至胆胰管中,为后续的活检、取石以及碎石等手术建立通道。为保障介入安全性,设计消化内镜机器人传感方案,实时检测内镜介入状态,包括输送力、输送长度以及输送速度等。(2) 针对消化内镜及镜下手术臂介入的需求,研究主从式介入方法。基于人机工程学设计建立消化内镜手术机器人与操作手柄的运动映射方案。针对腱鞘摩擦导致消化内镜及镜下手术臂控制精度低的问题,建立腱鞘驱动机构的背隙模型,并设计背隙补偿策略,提高手术臂控制精度,以完成精确的内镜弯曲定向以及手术臂的选择性插管任务。基于监督学习方法融合多源介入感知信息,对内镜介入操作进行辨识和异常检测,并设计内镜操作策略,为医生提供操作约束,降低主从操作模式下消化内镜的介入难度,保障介入安全性。(3) 针对消化内镜以及镜下手术臂自主介入的任务需求,研究消化内镜手术机器人自主介入方法。基于Faster-RCNN目标检测和数学形态学图像处理方法从镜下图像中提取消化道解剖特征和介入靶点。针对胆胰管介入过程中十二指肠乳头会被镜下手术臂遮挡的问题,基于Median Flow方法追踪十二指肠乳头在镜下图像中的位置。基于检测的介入状态和解剖特征位置设计消化内镜和镜下手术臂控制方法,实现消化内镜手术机器人的自主介入。(4) 开展仿真模型、动物以及临床实验研究来验证消化内镜和镜下手术臂的主从、自主介入方法的实用性和可行性。联合临床专家制定机器人操作流程,包括手术场景布置、备台/撤台流程以及机器人术中可能出现的风险及相应的对策。研究制定详细的实验方案,记录重要实验数据,评估主从介入和自主介入方法在成功率、耗时等指标方面的表现。实验结果表明,自主介入的表现如耗时、介入力等要优于新手医生,且基本持平于专业医生的主从操作。综上所述,本文从消化内镜介入手术面临的手术器械不灵巧导致介入难度大、多人协作操控内镜和器械效率低以及辐射危害生命健康等问题出发,针对消化内镜手术机器人构型设计、控制策略以及自主介入方法等关键技术开展了深入的研究。通过开展模型实验、动物实验以及临床实验验证所建立的消化内镜手术机器人的实用性。本文提出的消化内镜手术机器人构型能够完成消化内镜及镜下手术臂的介入操作任务需求。本文提出的主从介入方法能够辅助医生实现机器人化消化内镜操控及镜下手术臂的介入操作。其中,研究的腱鞘驱动机构的背隙模型能够提升执行臂的定位和定向精度;提出的操作约束方法能够降低内镜介入难度,保障手术安全性。本文提出的消化内镜及镜下手术臂自主介入方法能够将医生从繁重的操作任务中解放出来,将精力更多地集中在操作决策层面。此外,本文所提自主介入方法能够拓展到其他自然腔道如呼吸、泌尿等的软式内镜的植介入诊疗。
Other AbstractERCP (Endoscopic Retrograde CholangioPancreatography) surgery is known as the most difficult and risky (the bright pearl of the crown) surgery in digestive endoscopy because of the special surgical site and the organs interferences. At present, there are still some challenges in clinical ERCP surgery, which make it highly dependent on the skills and experiences of endoscopists. These challenges are mainly including the inflexibility of the surgical instrument leads to difficulty in manipulation, the difficulty of multi-human cooperation causes low surgical efficiency and serious radiation hazards to the life and health of endoscopists. Funded by the National Key Research and Development Program, this project conducts depth research on the structural design, control strategy, and autonomous intervention strategy of the robotic digestive endoscopy. The specific research contents are listed below: (1) Task requirements of ERCP surgery are analyzed, and the structural optimization design of the endoscopic operation robot is presented based on the previously developed robotic endoscopy system. The degrees of freedom of the endoscopic operation robot are configured by imitating the operation mode of clinical endoscopy to perform the robotic endoscopic intervention. Two continuum surgical manipulators are developed based on the structural and kinematic analysis, which can reach the duodenum along the instrument lumen of the duodenoscope. The proposed surgical manipulator can perform the precise cannulation to establish a surgical channel for biopsy, stone removal, and other operations. The sensing system of the robotic endoscopy system is designed to detect and display the interventional status in real-time, including insertion force, insertion length, and insertion velocity. (2) In response to the requirements for robotic endoscopic intervention and surgical manipulator intervention, the master-slave intervention method is investigated. Motion mapping between endoscope/surgical manipulator and joysticks is established based on ergonomics to realize the intuitive intervention. To face the problem of low positioning accuracy of the flexible manipulator (such as an endoscope or surgical manipulator) caused by the friction between tendon and sheath, a backlash modeling and compensation strategy based on a data-driven learning method is considered to improve the ability of positioning and orientation of tendon-sheath actuated flexible manipulator. A supervised learning method is proposed, which is based on the fusion of multi-information establish interventional situational awareness including interventional activities recognition and abnormal operation detection. A manipulating constraints strategy based on interventional situational awareness is designed to provide guidance and restraints for the endoscopist to reduce the operational difficulty in the master-slave control. (3) The autonomous intervention method of the proposed robotic endoscopy system is researched after analyzing the interventional task requirements. The anatomical features and intervention targets are extracted from the endoscopic images based on Faster-RCNN and the mathematical morphological image processing method. Aiming at the challenge that the papilla will be blocked by the surgical manipulator during the pancreatic-biliary duct intervention, the insertion position of the manipulator is tracked and extracted based on the Median Flow method. Control methods of the endoscope and surgical manipulator are designed based on the detected interventional status to realize the autonomous intervention. (4) Phantom, animal, and clinical experiments are performed to verify the practicability and feasibility of the master-slave intervention method and autonomous intervention method. The standard robotic intervention procedures are formulated in cooperation with clinical experts, including surgical scene arrangement, surgery preparation procedures, and corresponding solutions for complications during robotic surgery. Detailed experimental plans are researched and formulated, and the experimental results are recorded to evaluate the performance of master-slave intervention and autonomous intervention in terms of success rate, time-consuming, etc. Experimental results show that the performances of autonomous intervention are better than that of novice endoscopists, and can be comparable to professional endoscopists. In conclusion, to face the challenges of ERCP surgery, key technologies of robot structure design, control method, and autonomous intervention strategy are researched in this project. A series of phantom, animal, and clinical experiments are carried out to verify the practicability of the proposed robotic digestive endoscopy system. The experimental results show that the proposed structural design of the digestive endoscopy system can address the task requirements of the endoscope and surgical manipulator intervention. The proposed master-slave control method can assist endoscopist to perform robotic endoscope and surgical manipulator intervention, the presented backlash model of tendon-sheath driven mechanism can increase the positioning and orientation accuracy of the flexible manipulator, and the proposed manipulating constraints strategy can reduce the difficulty of endoscope intervention and improve the safety of operation. The proposed autonomous intervention method for endoscope and surgical manipulator can release the endoscopist from simple and repetitive tasks and make them concentrate more energy on decision making. In addition, the proposed autonomous intervention method can be applied in other natural orifices such as respiratory, urinary interventional diagnosis, and treatment.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/28999
Collection机器人学研究室
Affiliation中国科学院沈阳自动化研究所
Recommended Citation
GB/T 7714
江维. 消化内镜手术机器人关键技术研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2021.
Files in This Item:
File Name/Size DocType Version Access License
消化内镜手术机器人关键技术研究.pdf(34042KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[江维]'s Articles
Baidu academic
Similar articles in Baidu academic
[江维]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[江维]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.