1 / 18

Planning Among Movable Obstacles with Artificial Constraints

Planning Among Movable Obstacles with Artificial Constraints. by: Mike Stilman and James Kuffner. Presented by: Deborah Meduna and Michael Vitus. Motivation. http://www.cs.cmu.edu/~mstilman/mov/WAFR06-Planning.avi. Outline. Problem Definition Main Topics TRANSIT TRANSFER

remy
Download Presentation

Planning Among Movable Obstacles with Artificial Constraints

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Planning Among Movable Obstacles with Artificial Constraints by: Mike Stilman and James Kuffner Presented by: Deborah Meduna and Michael Vitus

  2. Motivation http://www.cs.cmu.edu/~mstilman/mov/WAFR06-Planning.avi

  3. Outline • Problem Definition • Main Topics • TRANSIT • TRANSFER • Algorithm Overview • Artificial Constraints • Obstacle Identification • Constraint Resolution • Results

  4. Problem Definition • Separate configuration spaces for different obstacle locations • Obstacle movement restricted to robot motion • Motion planning restricted to the robot subspace

  5. TRANSIT • Moves the robot along a path while obstacles remain fixed • Valid if and only if path is in the robot’s collision-free space

  6. TRANSFER • Moves the robot and ONE obstacle along a path to a new state • Valid if and only if: • The robot and obstacle paths are in collision-free space • The robot and obstacle do not collide along the path

  7. Problem Scope • Monotone vs. Non-monotone • Monotone: each movable obstacle only needs to be moved once • Non-monotone: can be broken into multiple monotone plans • Presented Planner is Linear-Monotone

  8. Algorithm Overview (1) • Plan a path through the cluttered environment • Allow translation through movable obstacles • Determine the last obstacle that has to be moved

  9. Algorithm Overview (2) • Plan a path to Transfer the last obstacle and Transit the robot to the goal • Adds artificial constraints for earlier timesteps • Resolve conflicts between movable obstacles and artificial constraints

  10. Artificial Constraints • Robot Transit operations create constraints on all obstacle configurations: • Robot motion along path sweeps volume V

  11. Artificial Constraints • Robot Transfer operations create constraints on non-moving obstacle configurations

  12. Reverse Search - Motivation • Assembly planning: • Much smaller branching factor due to actual constraints • Movable obstacles: • Final configuration not pre-determined • Must use forward search • Use reverse search for the ordering of which obstacles to move • Transfer of the last obstacle is performed first • Adds artificial constraints

  13. Obstacle Identification • Identifies last obstacle to be manipulated prior to reaching the goal or sub-goal • Utilize relaxed planner, Plast, allowing paths through movable obstacles • Select OL, last obstacle in collision with path

  14. Constraint Resolution • Plans a Transfer path for OL and the following Transit path to the goal • The two paths form artificial constraints • No obstacles scheduled earlier in time than OL can be within the two swept volumes

  15. Example • Show movie from CMU http://www.cs.cmu.edu/~mstilman/mov/WAFR06-Execution.avi

  16. Results • Figures 1, 2 and 4 not solve-able by existing planners 1 4 2 3

  17. Conclusions • Future Work: • Include accessibility constraints • Incorporate heuristics for generating Transfer paths

More Related