• Paper Planning Notes
• Paper Synopsis

Paper Outline

Bolded items represent paper sections and heading, grey items are things I'm not sure of

1. Background and Introduction
   • Motion planning refers to the process of finding an obstacle-free path for robots, given a starting point and a goal destination.
   • Applications of motion planning are wide ranging.
   • It is computationally hard to plan for robot navigation because of how difficult it is to capture all the robot constraints and represent the environment as a simple model to a computer.
   • Sampling-based Planning:
   • Our method propose a solution to the two bottlenecks of sampling-based planning
2. Related Work
   1. Rapidly-exploring Random Tree (RRT):
   2. Workspace Skeleton
      • Workspace skeletons are graphs which captures the topological features of the environment.
      • We use both the Reeb Graph and Mean Curvature Skeleton to generate our workspace skeleton. Mainly because, these skeleton are what's used for our skeleton-biased RRT, and because they both perform differently in workspace with different bodies.
      1. Reeb Graph
      2. Mean Curvature Skeleton
   3. Dynamic Region Rapidly-exploring Random Tree (DR-RRT)
      • DR-RRT is a skeleton-guided RRT which uses the workspace skeleton to bias RRT growth.
      • DR-RRT is faster and returns lower collision detection calls when compared to basic RRT
3. Our Method
   • Clearance-value is defined as the size of free space between obstacles in the environment
   • Our method is designed for skeleton-guided RRTs and applied to DR-RRT.
   1. Algorithm Overview
      • Detailed explanation of algorithm.
   2. Calculating Clearance-value
4. Experiments
   • We tested our method in robotics and protein environments.
     • I would also explain the experimental setup
   1. Robotics Experiments:
      • MazeTunnel:
      • Obstacles:
   2. Protein Experiments?
5. Discussion [ Experimental Analysis ]
   • I would explain the experimental analysis here
6. Conclusions
   • Our method utilizes clearance to improve RRT exploration in faster time with lower CDCs when compare to DR-RRT.
   • Our method returns safer paths when RRT is biased using maximum clearance
   1. Future Work
      • Compare our method to current methods in Image-guided Medical Needle Steering by running 3D experiments. The medical robotics application
      • Run our method is animation environments and compare with current methods for planning animation movements
7. Acknowledgments
8. References