Motion Planning for Laparoscopic Surgey
COMP 790-058: Motion Planning in Real and
Virtual Worlds
Course Project
Mert Sedef
Download Proposal Presentation
1. TOPIC
I am planning to work on the motion planning of a 4-dof laparoscopic surgical tool in the 3D abdominal region of human body to reach a desired goal point in the abdominal region.
Abdominal region of human body is a highly-dynamic environment which is filled with highly-deformable viscoelastic tissues and organs placed on top of each other with very little space. In the usual laparoscopic surgery settings, to be able to reach a desired goal location (like a tumor on an organ) in the abdominal region, the surgeon has to manipulate (deform and move) the organs on his way to the goal point using the laparoscopic tools. Based on experience, the surgeon implicitly knows how much force to apply to the organs he is deforming such that the deformed organs will not be harmed.
I would like to work on an algorithm that will plan the motion of the 4-dof surgical tool in such a scenario. Given the dynamically deformable abdominal region (which will include the dynamically deformable organs in situ, in their natural place in the human body), the robot should be able to plan a path to reach a target point in between these organs. While it tries to reach the target point, it will have to move the organs on the way by doing least amount of work and giving minimum harm to the organs, and be aware of some “penalty” tissues that should not be touched at all. Therefore, geometrically speaking, most of the time, there will not be a collision-free path from the starting point to the goal point. However, in this planning algorithm, the tool should find the shortest path and open it by pushing and deforming the other organs on the way to create a path. So, at this stage, the keywords here for this motion planning idea is “4-dof rigid robot”, “dynamically deformable environment”, “awareness of penalty tissues” "shortest path", "minimum work", "minimum tissue deformation-damage".
2. MOTIVATION:
One motivation is Robotic surgery. Robotic surgery can be classified into three stages: Preoperative surgery, Intraoperative surgery, and Postoperative surgery. As the name imply, Preoperative surgery is the part before the surgery where the flow of surgery is planned based on the patient-specific data. During this part, a motion planner algorithm can be used to find out the ultimate path of the surgical tool and the most efficient and least harmful maneuvers to follow during the Intraoperative part, which is the part where actual surgery takes place.
The other motivation is assessment of performance and training transfer on surgical simulators. There are a number of quantitative performance measures that are kept track of during a training session of a trainee on a surgical simulator to be able to asses the performance of the trainee on a virtual surgical task on the simulator. Some of these measures include task completion time, hand motion economy, path length, work done by trainee, and amount of unnecessary tissue damage. With a motion planner algorithm designed for a specific virtual surgical task, the optimum values of measures can be calculated and the values of a trainees performance can be compared with the optimum ones for a realistic and correct assessment.
3. PRIOR WORK:
Motion planning for deformable objects and environments is a very recent research area. Some of the past studies include
4. GOALS:
In addition to having a working motion planning algorithm such as the one described in Section 1, while working on the project, I would also like to come up with different novel ideas from the view point of robotics and motion planning that can be applied in the areas of Robotic Surgery and Surgical Simulation & Training, since these areas are pretty open to new ideas in these view points.
5. SUBTASKS:
The subtasks and sub-goals possibly cover finding out a reasonable motion planning technique that works under given constraints described in Section 1, applying a realistic physics-based modeling technique to model the deformations of the organs and tissues in the environment, and maybe trying to make the motion planning algorithm work at real-time update rates.
References