/mp500lwa4d_robot_env

This a mobile robot, compose of mobile base(neo_mp_500) and arm(lwa4d).

Primary LanguageC++

mp500lwa4d_robot_env

There are ROS packages of a mobile robot, consists of mobile base(neo_mp_500), arm(lwa4d) and gripper(PG-plus70), and a OpenAI.gym-based enviroment interface.

3dRobot

system

Before train the model, pleace determine the default version of Python you are using,

  • if it is Python2: you should install gym package in Python2
  • if it is Python3: you should install ROS in python3.6 enviroment. For how to install ROS kinetic in python3.6 enviroment, please see the document < Install ROS-Kinetic with Python3.md >

To train the model with a DDPG Agent:

  1. Set param ON_TRAIN in script main_distri.py line 19 and RLAgent.py line 12
ON_TRAIN = True
  1. Then run in bash
conda activate mp500lwa4d_gym
roslaunch mp500lwa4d_description RL_gazebo_distri.launch
rosrun robot_service RLAgent.py
rosrun robot_service main_distri.py