Research On Uncalibrated Visual Servo Control Based On A Binocular Magnetic Anchor Endoluminal Surgical Robot

Compared with traditional surgery, minimally invasive surgery has many advantages, like small trauma, less bleeding, short recovery period and low probability of complications. LaparoEndoscopic Single-site Surgery(LESS)is a kind of surgery that just need an incision to insert surgical instruments into human body, as it can bring better minimally invasive surgery effect. Now, LESS is successfully used in surgeries like cholecystotomy and appendicectomy. However, defects still exist, as surgical instruments of LESS are easily interfere with other devices because of the lack of triangle relationship. So, Intraluminal LESS robot can ease doctor’s fatigue and improve the quality of surgeries, but its control method often use open loop control mode. In the open loop control mode, operations are conducted under the visual feedback with low degree of automation and accuracy. To solve this problem, single-arch laparoscopic technology, robot technology, magnetic anchor technology and navigation technology are integrated in the paper, and research of uncalibrated visual servo control of magnetic anchored robot based on binocular vision is conducted.System scheme and binocular vision are designed according to the requirements of the surgical robot. To realize accurate control, visual servo method that uses image information to drive intraluminal surgery robot to target regions is adopted. In addition, binocular stereoscopic vision is used to obtain real-time surgical robot pose and precise location of the target point in the operation. According to requirements of object distance and regional vision in operations, parameters of lens are calculated to meet the need of field depth. With the method of depth error analysis, baseline is then designed to satisfy demands of the position measurement.Poses and locations of surgical robot and target points that would be used to control robot are resolved with features extracted from images. By establishing robot kinematics model and binocular camera model, 3-D coordinates of markers and joint vectors of robot can be easily obtained, and visual measurement can also be reached.In the condition that robot and visual unit have relative motion, uncalibrated visual servo control scheme that uses update law to estimate real-time velocity mapping relationship between joint space and pixel space of the robot is proposed, based on the characteristic of abdominal operation environment and robot’s decentralized configuration. Enhancement and update law’s influence on control results is analyzed by substituting parameters, and the feasibility of the control algorithm are verified by simulation.Finally, magnetic anchoring experiment platform is built, and the feasibility of visual servo control algorithm for magnetic anchored surgical robot is verified under the natural light. Moreover, experiments under simulated conditions of abdominal cavity are conducted, thus feature extraction algorithm’s stability and control algorithm’s stability in the environment close to real operations are verified.