Contact Us
CiiS Lab
Johns Hopkins University
112 Hackerman Hall
3400 N. Charles Street
Baltimore, MD 21218
Directions
Lab Director
Russell Taylor
127 Hackerman Hall
rht@jhu.edu
Last updated: 5/10/2023 1:00pm
We are designing an attachment to the Endo360 with STAR robot to autonomously manage the tensioning of the thread as the robot is suturing to allow for fully autonomous suturing with the STAR robot.
Figure 3. Diagram of project goal: small-scale mechanism which attaches to the Endo360 with STAR and can be inserted into the abdomen laparoscopically.
Increasing the level of autonomy in surgical robotic systems, particularly for time-consuming and repetitive tasks such as suturing, can help to standardize patient outcomes and free up surgeons to complete other tasks. Anastomosis (see figure 1) is a surgical procedure in which a luminal structure is reconstructed. It requires a high degree of maneuverability and repeatability and is thus a good candidate for examining autonomous robotic surgery on soft tissue [5].
The Endo360 with Smart Tissue Autonomous Robot (STAR) performs autonomous suturing in laparoscopic anastomoses. STAR has been shown to increase the consistency of suturing compared to both traditional manual surgery and non-autonomous robotic surgery [3]. However, STAR requires two incision sites to accommodate both laparoscopic arms, one for autonomously placing the sutures and one which is controlled by a human to manage the suture tensioning. Additionally, the need for human assistance in suture tensioning management means that STAR is not fully autonomous, creating a task which the surgeon must perform manually.
Therefore, the goal of this project is to develop an autonomous mechanism which enables STAR to perform single-arm robotic suture management and tensioning. This mechanism will replace the need for manual tensioning management and thus increase the level of autonomy of STAR. This will decrease the workload of the surgeon and also decrease the invasiveness of STAR procedures, since only one incision site would be needed as opposed to the two which are needed for the current dual-arm approach.
Figure 2. STAR current autonomous suturing workflow. Robot motion control and stitch placement is currently fully autonomous, but the assistant tool for suture tensioning management is controlled manually.
In order to accomplish this goal, our final mechanism must accomplish these specific design requirements:
Below we have outlined an overview of the tasks necessary in suture tensioning management. Our mechanism must be able to accomplish each of these tasks autonomously. We have included a design task which will be implemented into our mechanism to complete the given step of the workflow.
Workflow of suture management:
Our technical approach for this project is to first design and test a large-scale prototype for autonomous suturing outside of the body before moving forwards with the design of a small-scale prototype based on what we have learned from the initial prototyping.
Mechanical Design
For our first round of prototype development, we are creating a large-scale prototype which attaches to the arm of the KUKA robot. The purpose of this prototype is to test a mechanism which we plan to further miniaturize so that it is useful in laparoscopic surgeries. We will test the functionality of this prototype outside of the body and this will inform how we will move forward with our small-scale prototype.
Our large-scale prototype is mounted on the 7-DOF KUKA lightweight arm attached to the STAR and consists of two components: a swing mechanism at the front and a suture-tensioning rotor in the midsection. The “swing mechanism” catches the thread after the Endo360 completes a stitch and brings the thread to the midsection reel for tensioning.
Figure 4. Large-Scale Prototype CAD model with labeled components
a）Swing Mechanism (see figure 4) The “swing mechanism” catches the thread after the Endo360 completes a suture and brings the thread to the midsection reel for tensioning. The tip of the STAR is retrofit by a commercially-available manual laparoscopic suturing tool, Endo360. The Endo360 has 1-degree of freedom and uses a motorized circular needle to place sutures. As the stitching is a periodic movement, sutures can always be caught by the same side of the tip's caliper. The swing mechanism has a free-spinning rubber wheel, which is used to grasp the thread after the Endo360 completes a stitch. The swing mechanism catches the thread on its rubber wheel and then brings the thread to a rear roller, allowing the midsection reel to catch the thread.
b) Midsection roller (see figure 4) The midsection reel tensions the thread, which is driven by the rear motor through a belt pulley. The reel is designed with a curved hook which points in the rotation direction, allowing the reel to catch and tension the thread without tangling. When the suture is adequately tensioned, the rear motor will detect an increase in torque. This will tell the motor to stop and then start spinning in the opposite direction to release the thread.
Mechanical Design of Small-Scale Prototype
Figure 5. CAD model of gripper mechanism (left) and diagram of current design for small-scale prototype (right)
The small-scale version of this design will utilize a cable-driven gripper to grasp the thread after the STAR finishes a suture as shown in figure 5. This gripper will be mounted on a small linear actuation mechanism which will allow it to move out to catch the thread after the Endo360 completes a stitch while maintaining a small footprint, thus allowing the device to fit inside of a 25mm diameter hole. This prototype will maintain a scaled-down version of the “midsection roller” on the large-scale prototype for suture tensioning and will otherwise function in a similar manner to the large-scale prototype. This design is subject to change as we learn more through prototyping and testing.
Control System/STAR Interface
The whole structure is controlled and served by the rear-motor and catching motor, which are operated by ROS via CAN system. Besides, based on the position-model control, the motor could drive the swing mechanism to bring the suture thread to a specific place to tension it. For torque-model control, the motor could precisely detect whether the suture thread has already been tensioned. By communicating via the CAN system，Ros could control every motion and cooperate with the whole device with STAR robot so as to realize autonomous suturing operations.
Testing Plan
We will implement testing to ensure that our prototype can successfully replace the role of manual suture tensioning assistance currently utilized by the STAR system. Therefore, testing will consist of 2 phases: phase 1 mainly checks prototype functionality while phase 2 evaluates the performance of our device in comparison to the current dual-arm suturing approach.
During the first phase of testing, we will test the performance of each step of the autonomous suturing workflow process which we have described in the overview section.
Note that these 3 steps occur in both the large- and small-scale prototype. This allows us to compare the performance of the prototypes against one another. This testing is an important part of prototyping since it will provide valuable information about current design flaws.
The second testing phase focuses on the small-scale device and how its performance compares to the current STAR autonomous suturing workflow with the dual-arm setup. To simulate a suture in the human body, the prototype will be mounted onto a pre-existing test setup where it will suture synthetic skin. This allows us to analyze commonly used metrics in literature. Firstly, the total stitch time and number of failed sutures can now be measured more accurately since a full stitch procedure is performed. On top of that, the setup allows us to analyze the bite depth. This is important to check how consistent the device sutures and is a common analyzed metric. Lastly, we want to analyze the workspace the prototype occupies in a Solidworks motion analysis and compare this to the workspace of the dual-arm approach as a measure of the invasiveness of our device.
The first stage of this project consisted of making a dedicated design requirements document. This document influenced our design decisions to ensure that we are creating a product which actually solves our problem. Many of our design specifications constrained the size of the device to ensure that it could operate within a laparoscopic workspace. Some of the most notable form specifications are:
The next phase of the design process involved brainstorming and prototyping. After brainstorming many different possible solutions with the whole team, we decided to focus on physical prototyping of two designs which we felt were most likely to achieve our design specifications. Detailed design documents are linked below.
Figure 6. Roller Design
The prototype comprises two parts, one stationary and one swinging, and offers two degrees of freedom as depicted in Figure 3, where the Endo360 points downwards. Motor 1 controls the swinging arm, while motor 2 actuates the fixed rotor responsible for tightening the thread. The device functions by swinging the rotating arm from side to side, pulling the suture through the tissue in the process. The thread is squeezed between the swinging arm and the actuated rotor, while the end of the swinging arm spins passively. The power transmission would occur via a cable system. Note that the fixed and swinging rotors have a complementary shape optimizing the friction to tighten the suture. The team ultimately chose not to move forward with this design because of its large form factor and because it would require two motors, which is more complex than our final design.
Figure 7. Gripper Design
The gripper design is displayed in its non-actuated form in Figure 4.a. It works by spinning motor 2 which activates a cable mechanism that opposes the spring force holding the suture tensioning device in its closed configuration, causing it to move to its open configuration (Figure 4b). The thread will be caught between the two rollers on the gripper when the gripper is closed, one actuated and one free-spinning. In the closed position, motor 1 will actuate the driven roller, pushing all the loose thread towards the end effector until a certain current is detected on the motor, indicating that the sutures have adequate tension. At this point, motor 1 will stop moving, and motor 2 will actuate to return the gripper to its open configuration, allowing the STAR to place the next stitch. The team chose not to move forward with this design because of its large form factor, durability concerns with the springs, and because it again required two motors which is more complex than our final design.
Figure 8. Swing Mechanism Design (left) and transmission design (right)
The final decision was made based on which prototype would have most potential in meeting our design requirements. We decided to move forward with a swinging mechanism design because of its potential for a small form factor and the fact that it only required 1 DOF, and thus only one motor, simplifying our cable transmission system.
The swinging mechanism catches the thread and tensions it by rotating about a pin joint which causes it to pass by the STAR end effector. The motor actuating the lever arms is placed about 12” up the STAR shaft and transmits power through a cable connection. In the lever arms, the cables are guided by dedicated slots in the arms and fixed using screws. In the transmission part up the shaft, the cables have dedicated slots and guides on the motor spool. They terminate on endpoints by a press fit. The cables wind and unwind on a spool making the entire mechanism swing. Note that both cables are connected to the same motor. Returning to the default position is achieved using torsion springs located in dedicated grooves. These springs make for a straightforward transmission because it allows the mechanism to swing forwards and backwards using only one motor. Both the housing for the levers and the motor housing are fixed to the STAR system shaft using set screws.
See documents linked below for full testing procedures. Testing results are summarized in this section.
Figure 9: FEA results of the SMD: a) under normal tension force 0.6N b) under maximum tolerance with deformations.
Three rounds of testing were performed on the prototype. The first round of testing was a finite element analysis of our CAD model to verify that the prototype could withstand the forces necessary to complete the task. The analysis was carried out using SolidWorks software, taking into account the material properties of Dental Resin, the material which was used for 3D printing the prototype. The simulation defined the swing mechanism's side surface as a fixture, and gradually increased force applied to the swing mechanism tip. Both structures under normal stress and maximum stress were analyzed. The FFEPlus iterative solver was used to perform this analysis. The mesh density was kept moderate by setting the side length of the mesh triangle to 2.303mm. The maximum stress was primarily concentrated at the tension rotor's slot. The swing arm's maximum tolerable stress is about 1.315×10^7 N/m2, which means that it can withstand a maximum force of 13.4 N. Since the maximum tensioning force that this device must withstand is known to be about 1N [5], this analysis indicates that the prototype is sufficiently strong for our purposes.
Force Testing Results (left) and setup (right)
The force testing results indicate that our device in its current state is not capable of exerting a force of 1N without breaking. Sending 350mA current to the motor resulted in a 0.6N force on the force sensor. When stepping up to a 400mA current, the torsion springs on the pin joint slipped out of place causing a mechanical failure. Therefore, the maximum force that this prototype is capable of exerting on the suture is 0.6N.
At the time of testing 5/5/2023, the team was still waiting on bearings to arrive (which had been ordered four weeks prior). The team decided to go ahead with assembly and testing without these parts regardless. The team plans to incorporate bearings into the design when they arrive and run this test again at that time. The bearings will significantly strengthen the pin joint, which is where this mechanical failure occurred.
The purpose of the third round of testing was to verify that the prototype was capable of working with the existing STAR system to pull thread through synthetic bowel after the STAR places a stitch without interfering with the STAR system. We observed that our device was capable of exerting adequate force to pull suture through the synthetic bowel without interfering with the STAR end effector.
This is primarily a hardware project, so we will not be collaborating on code.
We will keep all CAD files and documentation in this google drive folder: https://drive.google.com/drive/folders/1FK375fwveAeCQHrSV9fCIeh-iuU8x9Ly?usp=sharing
