Joanneum Research Institute Release Version 1.0.0 Of Robo-Gym, An Open Source Toolkit For Distributed Deep Reinforcement Learning On Real And Simulated Robots

336
Source: https://github.com/jr-robotics/robo-gym

Deep Reinforcement Learning (DRL) has proven to be extremely effective when it comes to complex tasks in robotics. Most of the work done with DRL focuses on either applying it in simulation or using a real-world setup, but there are also examples that combine the two worlds by performing transfer learning. However, this approach requires additional time and effort because you need know how each system works individually before combining them together effectively. In order to increase the use of Deep Reinforcement Learning (DRL) with real robots and reduce the gap between simulation and robot control, Joanneum Research’s Institute for Robotics has released version 1.0.0 of robo-gym, an open-source framework that can be used by AI developers in developing reinforcement learning algorithms for controlling robotics devices more effectively than ever before.

The release was announced on the Robot Operating System (ROS) discussion forum. The release of a new version of the robo-gym framework makes it easier for researchers to develop their RL algorithms in simulated environments and transfer them to real world robots with minimal updates. The robo-gym is based on ROS, uses Gazebo as its physics engine, and offers OpenAI Gym interface support for simulation.

In 2020, Joanneum Research developed robo-gym and described the system in a paper presented at IROS ( International Conference on Intelligent Robots and Systems). The initial release of robo-gym ran on Python 3.5 with ROS Kinetic support for two separate robots: MiR100 mobile robot and UR10 collaborative industrial robot but now it has been updated so that it requires only python 3.6 or higher and ROS Melodic while dropping any previous versions from being supported like the ROS kinetics. It does include additional robot drivers for all Universal Robot models: UR3, UR3e, UR5, UR5e, UR10, UR10e, and UR16e.

Github: https://github.com/jr-robotics/robo-gym

Paper: https://arxiv.org/pdf/2007.02753.pdf

Tool: https://sites.google.com/view/robo-gym

Source: https://discourse.ros.org/t/robo-gym-v1-0-0-release-with-new-environments-and-noetic-support/20512

https://www.youtube.com/watch?v=XQmjf4Ktsvw&t=8s

A message from Asif Razzaq, Co-founder of Marktechpost:

Show your support for our mission ‘making AI understandable for all’ by joining/connecting through our 34k+ FB Group, LinkedIn Page and Quora AI Group.

Advertisement/Sponsored Post:

If you are a company looking to promote your product/webinar/conference/service, feel free to reach out via email to [email protected] We offer sponsored posts and advertisements.