Artificial Intelligence Meets Robotics at Robocup
Robotics, a branch of Electronics and Computer Science Engineering, is now being combined with Artificial Intelligence (AI) for various applications. One such application is in the game of soccer, where robots are connected via AI. This event is called Robocup, and researchers compete every year to showcase their robots at the Robocup Challenge.
Pyrus: A Python-Based Platform for Robocup Simulation
Researchers from Dalhousie University and Memorial University of Canada introduced Pyrus, a Python-based platform for simulating Robocup. They mentioned in their research paper that Pyrus makes it easy to train and test models. Currently, the common frameworks used in Robocup are HeliosBase and Cyrus2DBase, which are primarily built using C++. However, Pyrus has the advantage of being simple and accessible, making it suitable for beginners to test their models.
Tackling Noise with Reinforcement Learning and Machine Learning
The Robocup environment can be noisy, posing a challenge for robots. To address this issue, researchers implemented Reinforcement Learning and Machine Learning models for tasks like dribbling and passing. This helped reduce the noise and improved the robots’ performance.
Robocup has opened up opportunities for data enthusiasts to solve various Data Analytics problems. Pyrus is also being utilized to solve basic Machine Learning challenges related to Robocup. Researchers are continuously working on improving the Pyrus base code and plan to implement a Python monitor and log analysis software to enhance the model’s feasibility.
For more details, you can refer to the original research paper.
If you like our work, please follow us on Twitter.
About the Author: Bhoumik Mhatre is a third-year UG student at IIT Kharagpur, pursuing a B.tech + M.Tech program in Mining Engineering with a minor in economics. He is a Data Enthusiast and currently doing a research internship at the National University of Singapore. He is also a partner at Digiaxx Company.