FRobs_RL Resources
Collection of resource files for FRobs_RL. Includes object meshes, robot descriptions, MoveIt configurations, task enviroments, training logs, trained models, and more.
Robots resource repositories
List of all robot resources, their source repositories (if available), and what commit the resource was taken from.
| Robot or Resource | Source | Commit |
|---|---|---|
| ABB120 | https://github.com/ros-industrial/abb_experimental | d41bfe3af65b25c4860ec495fd8ccc31e0585480 |
| ABB140 | https://github.com/FreddyMartinez/abb_irb140_support | 417b4f826e59b1f744bcb40b6a0ce7c4d9189dfd |
| UR5 | https://github.com/ros-industrial/universal_robot | 06d8b9e2f5f86aa54f9f2845f11edbc84e2f951e |
| Kobuki | https://github.com/yujinrobot/kobuki | 23748ed3dfb082831ca8eaaef1a0b08588dbcb65 |
Enviroments list
List of all enviroments with action and observation space.
| Name | Description | Observation Space | Action Space |
|---|---|---|---|
 ABBIRB120ReacherEnv-v0 |
An enviroment where a manipulator robot needs to reach a spatial position. | Box(12,) | Box(6,) |
 ABBIRB140ReacherPosEnv-v0 |
An enviroment where a manipulator robot needs to reach a spatial position. This enviroment uses joint positions actions. | Box(12,) | Box(6,) |
 ABBIRB140VelReacherEnv-v0 |
An enviroment where a manipulator robot needs to reach a spatial position. This enviroment uses joint velocities actions. | Box(12,) | Box(6,) |
 UR5ReacherEnv-v0 |
An enviroment where a manipulator robot needs to reach a spatial position. | Box(12,) | Box(6,) |
 KobukiEmptyEnv-v0 |
A simply enviroment where a Kobuki robot needs to get to a static random position in an empty enviroment. | Box(2,) | Box(2,) |
 KobukiDynamicEnv-v3 |
A slightly more complex version of the kobuki enviroment where the robot needs to get to a random position avoiding eight boxes with random velocities and trajectories. To identify the boxes the robots has a LIDAR sensor with 36 rays. | Box(38,) | Box(2,) |
 KobukiMazeEnv-v0 |
The more complex kobuki enviroment where the robot needs to follow 4 discretized trajectories in a maze with dynamic objetcs. To identify the boxes the robot has a LIDAR sensor with 8 rays. To follow the trajectory a midway point reward is used. | Box(15,) | Box(2,) |