RAMP header logo
Demo Program


Integrated Motion Planning Components


The MPK system may be used to evaluate the performance of a wide array of motion planning algorithms. Researchers will develop their own algorithms within the MPK framework and add them to the system. In current version of MPK, various planning algorithms have been implemented for different planning problems.

For basic motion planning problems, where the start and goal configurations are given, the following planning algorithms have been implemented:
  • ACA: Ariadne's Clew Algorithm [1]
  • RRT: Rapidly-exploring Random Trees [2]
  • RRT-Connect [3]
  • PRM: Probabilistic Roadmap Method [4]
  • Lazy-PRM [5]
  • SANDROS [6]
  • Sequential Framework [7]
For trajectory tracking problems, where a robot manipulator is required to move along a given end-effector path/trajectory, the following planning algorithms have been implemented:
  • Jacobian-based pseudo-inversed method [8]
  • Probabilistic Method [9]
For robot manipulator inverse kinematics problems, where the start configuration and the goal end-effector pose (position and orientation) are given, the following algorithm has been implemented:
  • Kinematics roadmap [10]
For closed-chain robots planning problems, the following algorithm has been implemented.
  • Randomized Gradient Descent method [11]
  • Active-passive Link Decomposition [12]
For path planning problem with end-effector constraints, the following planning algorithms have been implemented:
  • Adapted-RGD [13]
  • ATACE: Alternate Task-space and C-space Exploration [13]

Collision Detectors

The most desirable manner of supporting collision detection is to permit interfacing with as many existing packages as possible and to allow the programmer or the user of the interactive system to choose the one that he or she feels is appropriate. The MPK can be used to evaluate the utility of different collision detection schemes in much the same way as it is used to evaluate the performance of motion planning algorithms.

Collision detectors that are currently integrated into the system include:
  • A homegrown, simple collision detector,
  • V-collide [14]
  • Swift++ [14]
  • Solid [15]

Path Smoother

More introduction/reference to be added.

Updated on May 10, 2007 by Zhenwang Yao .
SFU logo
RAMP Lab vertical logo