6.3.2.4. Robot Manipulator on Homogeneous Matrix
Ver. 1.1.9 (2023-08-21)
This module provides an environment of a robot manipulator based on Homogeneous Matrix
This environment represents the robot manipulator in term of mathematical equations. The mathematical equations are based on rigid body transformation. In this case, the Homogeneous Transformation Matrix (HTM) is used for the structure. HTM is a matrix that contains both the translation rotation of a point with respect to some plane.
This robotinhtm environment can be imported via:
from mlpro.rl.pool.envs.robotinhtm import RobotHTM
Prerequisites
General Information
Parameter |
Value |
---|---|
Agents |
1 |
Native Source |
MLPro |
Action Space Dimension |
[4,] |
Action Space Base Set |
Real number |
Action Space Boundaries |
[-pi,pi] |
State Space Dimension |
[6,] |
State Space Base Set |
Real number |
State Space Boundaries |
[-inf,inf] |
Reward Structure |
Overall reward |
Action Space
By default, there are 4 action in this environment. The action space represents the angular velocity of each joint of the robot manipulator.
State Space
The state space consists of end-effector positions (x,y,z) of the robot manipulator and target positions (x,y,z).
Reward Structure
By default, the reward structures are shown in the following equation:
- Cross Reference
Howto RL-ENV-002: SB3 Policy on RobotHTM Environment
Citation
If you apply this environment in your research or work, please cite us and the original paper.