Open source toolkit eases reinforcement learning for robots

Open source toolkit eases reinforcement learning for robots
New Products |
Acutronic Robotics has designed a toolkit for developing and comparing reinforcement learning algorithms using ROS 2 and Gazebo.
By eeNews Europe

Share:

Built as an extension of gym-gazebo, gym-gazebo2 has been redesigned with community feedback and adopts now a standalone architecture while mantaining the core concepts of previous work inspired originally by the OpenAI gym.

The company describes gym-gazebo2 as a complex piece of software for roboticists that puts together simulation tools, robot middlewares (ROS, ROS 2), machine learning and reinforcement learning techniques. All together to create an environment where to benchmark and develop behaviors with robots. Setting up gym-gazebo2 appropriately requires relevant familiarity with these tools. Open code is available in Github, while in-depth explanations and actively growing tutorials can be found at https://acutronicrobotics.com/docs

“This is the logical evolution towards our initial goal: to bring RL methods into robotics at a professional and industrial level”, explains Risto Kojcev, Head of AI at Acutronic Robotics. The AI team he leads researches on how RL can be used instead of traditional path planning techniques.

“We aim to train behaviours that can be applied in complex dynamic environments, which resemble the new demands of agile production and human robot collaboration scenarios”.


Achieving this would lead to faster and easier development of robotic applications and moving the RL techniques from a research setting to a production environment. gym-gazebo2 is a step forward in this long term goal. In a paper available here, the company presents the new ROS 2 based software architecture and summarizes the results obtained using Proximal Policy Optimization (PPO).

The team has focused on MARA, a modular robotic arm that is natively running ROS 2 in each of its modules. They have evaluated four different environments with different levels of complexity of MARA, reaching accuracies in the millimeter scale. The environments are MARA, MARA Orient, MARA Collision and MARA Collision Orient.

Get the open code and extra resources at: https://github.com/AcutronicRobotics/gym-gazebo2

Acutronic Robotics – https://acutronicrobotics.com

Linked Articles
eeNews Europe
10s