Georgia tech team unveils autonomous tennis robot to help you beat your foes

“It can also be my opponent. It can help me train."
Amal Jos Chacko
An image of ESTHER mid-shot.

An assistant professor of Interactive Computing at the Georgia Institute of Technology has revealed a robotic tennis partner that may soon become your sparring partner and skilled opponent.

Dr. Matthew Gombolay envisions a future where human-scale robots play a crucial role in sports and athletic training. His latest creation, ESTHER (Experimental Sport Tennis Wheelchair Robot) is inspired by the limitations of traditional static ball machines used for tennis training.

Unlike these conventional ball launchers, ESTHER is designed to simulate human opponents and thus, real match conditions, allowing athletes to improve their skills and performance more effectively.

A tennis enthusiast himself, Dr. Gombolay understands the need for a robot that could adapt to various playing styles and exploit weaknesses in a player's game. By leveraging the design of wheelchairs used in wheelchair tennis, Dr. Gombolay and his team solved the question of moving ESTHER over the court.

A breakthrough in human-scale robotics

Dr. Gombolay devoted two years to building ESTHER with the collaboration of over 20 students. The team's efforts led to an exciting breakthrough, resulting in the successful programming of ESTHER to locate an incoming tennis ball and execute a return shot consistently. While the robot falls short of matching the skill level of its namesake, the renowned wheelchair tennis player Esther Vergeer, the ability to build a human-scale robot capable of hitting a return shot is a remarkable achievement.

Two DC motors paired to a gearbox provide ESTHER with the agility and thrust to cover both sides of a tennis court. The main challenge lies in pathfinding, as the robot must anticipate the ball's trajectory and determine the optimal path to intercept it.

To address this, the team employs a network of high-resolution cameras positioned around the court, utilizing computer vision algorithms to enable recognition of the incoming ball. Multiple camera angles allow the triangulation of the ball's position in space, predicting its trajectory and determining the appropriate response.

Future applications and the potential for training revolution

Although ESTHER's current capabilities are limited to hitting back-and-forth rallies, Dr. Gombolay and his team have ambitious plans for the robot's future development. The next phase involves teaching ESTHER to strategize shot selection, elevate its playing abilities and enhance its value as a training tool.

By incorporating reinforcement learning methods, the robot would be capable of autonomously improving its decision-making and shot execution, becoming more aggressive and effective in winning games.

The implications of ESTHER's technology extend far beyond tennis, according to Zulfiqar Zaidi, a lead student on the project. “While tennis is a great starting point, a system that can play tennis well can have applications in other fields that similarly require fast dynamic movements, accurate perception, and the ability to safely move around humans.”

“This technology could be useful in manufacturing, construction, or any other field that requires a robot to interact with humans while performing fast and precise movements,” he added.

The revolutionary aspect of ESTHER lies in its potential to transform athletic training and preparation. Athletes could hone their skills by playing against the robot which would mimic the styles and tactics of specific opponents.

The day may not be far when robots like ESTHER become integral partners for athletes across various sports, pushing the boundaries of performance and redefining how we train.

Add Interesting Engineering to your Google News feed.
Add Interesting Engineering to your Google News feed.
message circleSHOW COMMENT (1)chevron
Job Board