Some path planning demos

These movies were generated using the flc format and/or the MNG format. To view the flc movies, use the xanim. To find viewers for the MNG movies, see here. The advantage of the MNG format is that the movies tend to be smaller than the flc movies.

The color of the robot is green at each configuration of the sequence of configurations of the path returned by the planner. The color is yellow when interpolating between the steps in the path. Unless otherwise specified, the robots do not have joint limits.

Smoothed paths

• Robot with 3 joints, four obstacles, 2048-node roadmap flc
• Robot with 6 joints, four obstacles, 8192-node roadmap flc
• Robot with 10 joints, 9 joints with limits, corridor obstacle, 8192-node roadmap flc MNG
• Robot with 19 joints, two obstacles, 2048-node roadmap flc

These plans were generated from a roadmap built with a new workspace-based distance function. All of these examples use a robot with 10 joints, 2048-node roadmap.

• Robot without joint limits, corridor obstacle: node-based smoothing only (flc), plus (171 sec) interpolated smoothing (flc MNG).
• Robot with joint limits, corridor obstacle: node-based smoothing only (flc MNG), plus (10 sec) interpolated smoothing (flc MNG).
• Robot without joint limits, four obstacles: node-based smoothing only (flc), plus (13 sec) interpolated smoothing (flc MNG).

Robots with 3D workspace

Finally, an example for a robot with a 3D workspace. The robot has six joints, and the roadmap has 2048 nodes and was created using the new workspace-based distance function used above.

• Cage obstacle: (23 sec) node-based smoothing only (flc); plus (5 min) interpolated smoothing (flc)

Next innovation: geometric description of obstacles (as opposed to listing each cell occupied by the obstacles individually).

• Cage obstacle, 3D workspace, 6 DOF robot: flc

Mobile Robots

If the base of the robot is translated along the X or Y axis by one grid unit, the obstacles look exactly the same, but shifted by the same amount. A planner was developed to take advantage of this. For this planner, a roadmap is constructed for the robot for all of its degrees of freedom except X and Y translation.

• "#" obstacles, 3D workspace, 6 DOF plus X and Y translation on the robot base: flc. Planning time: 78 seconds.
• "#" obstacles, 3D workspace, "L" shape: flc. The three orientation degrees of freedom and translation along the Z axis are part of the roadmap constructed for this robot. The X and Y translations are separate. Planning time: 2 seconds.
• "Forest" obstacles, 3D workspace, 6 DOF plus X and Y translation, example 1: flc. Planning time: 156 seconds.
• "Forest" obstacles, 3D workspace, 6 DOF plus X and Y translation, example 2: flc, smoothed flc. Planning time: 132 seconds.

Extras

• Corridor obstacle, planar robot, 15 joints: flc
.
• Four cages, 3D workspace, 15 joints: flc
• Forest obstacle, 3D workspace, 9 joint mobile robot: flc
• Forest obstacle, planar mobilr robot, 10 joints: flc
• "#" obstacles, "C" shape mobile robot: flc

Bad Paths

Examples of bad paths that can appear in the roadmaps.

• Straight-line path in configuration space resulting in a double spin in the workspace (flc, MNG). This robot has 20 joints, no joint limits, 16k nodes in the roadmap.