Opponent Hitting Behavior Prediction and Ball Location Control for a Table Tennis Robot

Area of Science:

• Robotics
• Artificial Intelligence
• Sports Technology

Background:

• Table tennis robot control often overlooks opponent interaction, limiting hitting accuracy.
• Existing systems lack adaptability to diverse opponent playing styles.

Purpose of the Study:

• To develop an advanced table tennis robot framework that incorporates opponent hitting behavior analysis.
• To enhance ball return accuracy and robot adaptability in table tennis.

Main Methods:

• Classifying opponent strokes into four categories: forehand attacking, forehand rubbing, backhand attacking, and backhand rubbing.
• Designing a robot arm with a two-dimensional slide rail for an expanded workspace.
• Integrating a visual module for capturing opponent motion sequences.
• Utilizing quintic polynomial trajectory planning for smooth robot motion control based on predicted ball trajectories.

Main Results:

• The proposed framework demonstrates effective integration of opponent behavior analysis for ball return.
• The specialized mechanical structure and visual module enable comprehensive motion capture and workspace coverage.
• Quintic polynomial trajectory planning ensures smooth and stable robot hitting control.
• Experimental results validate the strategy's effectiveness in achieving accurate ball returns.

Conclusions:

• The novel framework significantly improves table tennis robot performance by considering opponent behaviors.
• The system offers a robust solution for adaptive and accurate ball returns in robotic table tennis.
• This research advances the field of human-robot interaction in dynamic sports environments.