Google’s DeepMind Robotics team has reported making significant strides in robotics with its newly developed table tennis robot, as detailed in their paper, “Achieving Human Level Competitive Robot Table Tennis.”
This robot, designed to benchmark robot arms, demonstrates amateur human-level play against human opponents.
During tests, the robot consistently defeated beginner-level players and won 55% of matches against intermediate players. However, it struggled against advanced players, losing all matches.
Overall, the system achieved a 45% win rate out of 29 games played.
The research highlights the robot’s ability to play a sport with humans at a human level as a milestone in robot learning and control.
Despite this progress, the team acknowledges that achieving consistent human-level performance and developing robots capable of performing various real-world tasks remain significant challenges.
The primary limitations include the robot’s reaction to fast balls, system latency, and the need for advanced control algorithms and hardware optimisations.
Future improvements may involve predictive models and faster communication protocols to enhance the robot’s responsiveness and accuracy.
