Multi Geometry Robot

This Multi Geometry Robot environment can be installed via:

import mlpro.rl.pool.envs.multigeorobot

• 3rd Party Dependencies

  The environment has been tested in Ubuntu 20.04 running ROS Noetic.

  The installation steps are as follow:

  1. Install Ubuntu 20.04

  2. Install ROS

  3. Install Moveit

  4. Install Dependencies:

     sudo apt-get install ros-$ROS_DISTRO-gripper-action-controller ros-$ROS_DISTRO-joint-trajectory-controller
     sudo apt install ros-$ROS_DISTRO-moveit-resources-prbt-moveit-config
     sudo apt install ros-$ROS_DISTRO-pilz-industrial-motion-planner
     sudo apt install python3-pip
     pip3 install catkin_tools gym empy defusedxml pymodbus numpy netifaces pycryptodomex
     

  5. Build the Environment:

     cd MLPro/src/mlpro/rl/pool/envs/multigeorobot/src
     cd .. && catkin_make
     

  6. Source the package:

     echo "source MLPro/src/mlpro/rl/pool/envs/multigeorobot/devel/setup.bash" >> ~/.bashrc
     source ~/.bashrc
     

  7. Change the ros_ws_abspath parameter in:

     MLPro/src/mlpro/rl/pool/envs/multigeorobot/src/multi_geo_robot_rl/multi_geo_robot_training/config/multi_geo_robot.yaml
     

• Overview

• General information

Building a robot model in URDF format can be a painful work, if you do not know what you are doing. Especially when the model is complex. Multi Geometry Robot environment provides an easy way to build your own robot with a predefined configuration. This way, the user do not have to take care of the robot model. All the configurations are automatically built. The robot is also attached with a robotiq gripper. The robot can be configured in:

MLPro/src/mlpro/rl/pool/envs/multigeorobot/src/multi_geo_robot_rl/multi_geo_robot_training/config/multi_geo_robot.yaml

Below are the parameters on above mentioned file that can be configured:

Parameter	Description	Example value
robot_type	Type of the robot, “2D” or “3D”	“3D”
arm_num	Number of arm, positive integer value	3
arm_joint_seq	Arm Joint Sequence per arm, 0=Fixed Joint, 1=Revolute Joint	[[1,1,0],[0,1,0],[0,1,0]]
arm_length	Length of each Arm, list of positive floating value	[0.2, 0.2, 0.2]
arm_mass	Mass of each arm, list of positive floating value	[9, 2, 2]
adapter_mass	Mass of Adapter (connection between arm), positive floating value	5
eef_dummy	Display dummy ball on the end effector point, boolean value	False
max_iterations	Max cycle per episode, positive integer value	20
task_and_robot_environment_name	The name of the environment task registered on gym	‘MultiGeoRobotReach-v0’

For the joint sequence explanation is shown on the figure below.

The configuration will automatically build the robot model based on the above configuration. Below are the general parameters information regarding the Multi Geometry Robot environment.

Parameter	Value
Agents	1
Native Source	MLPro
Action Space Dimension	[predefined by the configuration,]
Action Space Base Set	Real number
Action Space Boundaries	[-0.1, 0.1]
State Space Dimension	[6,]
State Space Base Set	Real number
State Space Boundaries	[-2.0, 2.0]
Reward Structure	Overall reward

• Action space

The action of the agent directly affects the joint angles (rad) of the robot. The action is interpreted as increments towards the current value. The number of action depends on above configuration.

• State space

The state space consists of position information of the end effector (Red Ball) and the target location (Blue Ball).

Element	Parameter	Boundaries
PositionX	m	[-2.0, 2.0]
PositionY	m	[-2.0, 2.0]
PositionZ	m	[-2.0, 2.0]
Targetx	m	[-2.0, 2.0]
Targety	m	[-2.0, 2.0]
Targetz	m	[-2.0, 2.0]

• Reward structure

distance = np.linalg.norm(np.array(observations[:3]) - np.array(observations[3:]))
ratio = distance/self.init_distance
reward = -np.ones(1)*ratio
reward = reward - 10e-3

if done:
    reward += self.reached_goal_reward

• Version structure

  • Version 1.4.0