可撓式腹腔鏡使用者友善操作系統於機器人微創手術之應用

Abstract

腹腔鏡手術的應用已遍及許多手術的範疇。在過去，此種手術不僅需要反覆的訓練，也要求醫師與助手之間的高配合度與良好的默契。因此，人們致力於開發腹腔鏡手術機器人系統，藉由手術機器人的穩定性及易操作、不易抖動的特性，得以實現內視鏡自動操作的功能。 我們建立了機器人多關節可撓式腹腔鏡系統（RFLS），使醫師能藉由頭部控制操作腹腔鏡。然而，將機器人引進手術室，不僅需要考慮安全性，更需要考量使用者友善的議題，提高手術執行效率，且能夠真正將機器人與醫療結合。因此本研究主要是針對使用者友善的議題做探討，分為三大類：系統參數適應性、手術操作位置之記憶與取回、腹腔異物質偵測系統。系統參數適應性方面，針對因不同醫師對於機器人操作的熟悉度不同，系統會偵測使用者的操作流暢度，改變系統增益，使醫師能更加適應本系統。另外，外科醫師在手術操作時，常需使用腹腔鏡往返兩點之間的視野。針對此需求，我們提出手術操作位置之記憶與取回的功能，使醫師能以簡單的方式記憶當下的位置，並且能輕易地回到該次記錄的位置，此為腹腔鏡機器人之一大優點。進行微創手術時，常需尋找腹腔中的異物質(例如：腫瘤、息肉等)，因此我們提出了一套腹腔異物質偵測系統，在腹腔鏡的視野下，系統可自動辨別是否為異物質，且自動將畫面移動至異物質所在的中心點位置，幫助醫師在手術中能更方便地使用腹腔鏡機器人。實驗結果顯示自動參數調整可使得每位使用者能以最短的時間熟悉本系統，且此演算法平均可降低19.39%使用者完成手術操作的時間；記錄視野位置的準確度能提升手術流暢度，其取回的最終位置與記錄的位置誤差小於0.15度；腹腔異物質偵測系統能快速偵測異物的位置，且在3秒內準確且穩定地將畫面移動至目標處。

The practice of laparoscopic surgery continues to gain widespread adoption for a variety of surgical procedures. In the past, this kind of surgery required not only thorough training but also highly coordinated fashion between the surgeon and assistant. There-fore, people have done a lot of research on laparoscopic surgery robotic systems to fulfill laparoscopic endoscope navigation. We use a Robotic Flexible Laparoscope System (RFLS) to provide surgeons manipulate the laparoscope with his/her head movement. However, we need to concern not only the safety issues but also the user-friendly aspect if robot-assisted system is introduced into the practical surgical operation. It needs to improve the efficiency of the surgery, and be perfectly combined with surgical procedures. First of all, because the operational sophistication of surgeons is different, not every user is familiar with our RFLS. An adaptive intelligent control algorithm is integrated with our system so that the system will automatically adjust the parameters with every user based on their sophistication with the operation. Therefore, surgeons can get used to our robot-assisted system in a relatively short time. Surgeons usually need to manipulate the laparoscopic view back and forth between two targets during the surgery. We propose the Anatomical Position Retrieving Ability to record the anatomical position which the surgeon operates at the time. Surgeons can easily retrieve the previous position by a single command with small errors. Besides, the system can determine whether there is any tissue obstacle, and decide to retrieve the recorded position along a straight trajectory or previous one. In minimally invasive surgery, surgeons often need to find the foreign bodies inside the abdomen (i.e. tumors, polyps... etc.). As a result, we propose the Target Detecting System that identifies foreign bodies and moves the flexible laparoscope to the target. Experimental results show that simultaneous self-calibration will always work during operation. The algorithm can simultaneously adjust the system parameters with every user. It can reduce the task time up to 19.39% on average, making the overall system more user-friendly with the operator. Surgeons can accurately retrieve the previous position by using the Anatomical Position Retrieving Functions. The error between the recorded position and the end position is less than 0.15 degrees. The Target Detecting System can fast detect the target inside the abdomen and stably moves to the accurate target within 3 seconds.