HITTER: A HumanoId Table Tennis Robot via Hierarchical Planning and Learning

★

☆

SHARE

Overview

Paper Summary

Conflicts of Interest

Identified Weaknesses

Rating Explanation

Good to know

Topic Hierarchy

File Information

Paper Summary

Paperzilla title

Robot Learns to Play Ping Pong (But Needs Fancy Cameras)

Researchers developed a humanoid robot capable of playing table tennis using a combination of motion capture, planning algorithms, and learned control policies. While the robot successfully rallies with humans and other robots, its performance is limited by the need for motion capture and a simplified stroke set. It also struggles with short or deep shots due to a fixed hitting plane.

Possible Conflicts of Interest

None identified

Identified Weaknesses

Dependence on motion capture

The robot relies on motion capture cameras to track the ball and its own movements. This means it can't play in a normal setting without these specialized cameras.

Limited stroke repertoire and spin handling

The robot can only perform basic forehand and backhand hits, and can't handle spin. Real table tennis involves much more complex strokes and spin control.

Fixed hitting plane

The robot's hitting plane is fixed, limiting its ability to handle short or deep shots. A human could easily exploit this weakness.

Rating Explanation

This paper presents a well-executed robotics project with impressive real-world results. The hierarchical control system and integration of model-based planning with reinforcement learning are notable strengths. However, the dependence on external motion capture and limitations in stroke repertoire prevent a higher rating.

Good to know

This is our free standard analysis. Paperzilla Pro fact-checks every citation, researches author backgrounds and funding sources, and uses advanced AI reasoning for more thorough insights.

Explore Pro →