3D Benchmark Results

 

Overview

This page contains a comparison of two planning techniques, Uniform and Voronoi Based.  Uniform sampling refers to an implementation of a probabilistic roadmap planner using a uniform sampling approach.  The Voronoi Based Planner is identical to the Uniform sampler in every way, except that sampling is biased toward the medial axis using Hardware accelerated voronoi computations.

o        Result / Timings

o        Images / Descriptions

Results

  • RED indicates failure to complete in allocated time.
  • For a description of the columns in the table, click here.
  • Benchmark 4: A 3D environment with a straight tunnel. (a),(c) do not require rotation. (c) requires rotation.
  • Benchmark 5: A 3D environment with a "spiral tunnel". (a),(c) do not require rotation. (c) requires rotation.
  • Benchmark 6: A 3D environment with attempting to guide a piano through a window.

 

Bench

DOF

Type

Time

#C-Space samples

#Free samples

#Full Coll.

#QR Coll.

#Connect Comp. Calls

4(a)

3

Uniform

8.28

4669

2714

10148

0

375736

4(a)

3

Voronoi

3.48

471

378

3098

2908

12685

4(b)

6

Uniform

7892.55

126189

15079

212859

0

55216398

4(b)

6

Voronoi

721.28

35328

2707

33344

21088

1803603

4(c)

3

Uniform

14179.1

313014

54593

488821

0

57429630

4(c)

3

Voronoi

2956.83

93723

12718

93528

67087

4634587

5(a)

3

Uniform

296.65

61675

10985

98704

0

2028713

5(a)

3

Voronoi

160.99

22021

6087

52956

32091

715593

5(b)

6

Uniform

19223.3

733663

13335

845172

0

14035450

5(b)

6

Voronoi

15434.1

543401

9874

618233

152987

10329874

5(c)

3

Uniform

297.14

61675

5428

98704

0

2028713

5(c)

3

Voronoi

12.45

5672

472

2732

982

Not Applicable

6

6

Uniform

3948.86

91264

43542

250043

0

452459428

6

6

Voronoi

2983.24

58265

38098

220023

10821

351247399

                                                                                               

Benchmark 4 - 3D environment requiring the robot (a block) to move from one open are to the other by traversing a single channel. The channel provides for a narrow passage that is narrow mostly in workspace.

Tunnel Diameter is: .1 on a side.
Scenario (a) - Robot Size: .025 x .025 x .025, no rotation required.
Scenario (b) - Robot Size: .025 x .025 x .125, rotation is required.
Scenario (c) - Robot Size: .08  x .08  x .08, no rotation required.

 

Benchmark 5 - A 3D environment similar to the one in Benchmark 4, but much more difficult. The robot (again a simple block) must navigate through a single channel that spirals through the workspace. The environment is more difficult because the "elbows" cause multiple tight passages in configuration space.

Tunnel Diameter is: .1 on a side.
Scenario (a) - Robot Size: .02 x .02 x .125, rotation is required.
Scenario (b) - Robot Size: .07 x .07 x .07, no rotation required.
Scenario (c) - The environment is the same as in (a) except we use a different connection scheme to better use the information provided by the medial axis.

 

 

Benchmark 6 - A 3D environment with attempting to guide a piano through a window..

Scenario (a) - Piano must move from red to blue, moving through the window.
 
 

 

 

Copyright 1999. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the author.

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

geom@cs.unc.edu
last updated: 12/12/1999