Your data. Your choice.

If you select «Essential cookies only», we’ll use cookies and similar technologies to collect information about your device and how you use our website. We need this information to allow you to log in securely and use basic functions such as the shopping cart.

By accepting all cookies, you’re allowing us to use this data to show you personalised offers, improve our website, and display targeted adverts on our website and on other websites or apps. Some data may also be shared with third parties and advertising partners as part of this process.

RAI Institute / YouTube
News + Trends

Throwing, catching, hitting: This robot plays baseball

Kim Muntinga
13-10-2026
Translation: machine translated

With particularly flexible arms, the Robotics and AI Institute aims to demonstrate how robots can combine fast movements with precise force control. AthenaZero uses a special technical approach for this.

A baseball flies toward the robot. It aligns its arm, catches the ball, and throws it back. What looks simple poses major challenges for robots. They must react quickly and precisely adjust their force upon contact with the ball.

To this end, the US-based Robotics and AI Institute (RAI) has developed the research robot AthenaZero. The institute first demonstrated baseball exercises in 2025. In September, the researchers published their results in the journal «Science Robotics».

Two arms with minimal resistance

AthenaZero is approximately 1.60 meters tall and has two arms with a span of 1.80 meters. In total, the stationary robot has 27 degrees of freedom. These describe how many independent movements its joints allow. Of these, seven are per arm, six per hand, and one for the upper body. 22 motor drives control the movements.

The design of the arms is crucial. Conventional industrial robots often work with high gear ratios. In this setup, the motor rotates multiple times while the driven joint rotates only once. This increases the available force but makes rapid changes in direction and compliant movements upon contact with objects more difficult.

RAI is pursuing a different approach. The researchers use specially developed motor drives with low gear ratios. Furthermore, heavy components are positioned as close to the upper body as possible. This reduces the inertia of the arms. They can be accelerated and decelerated more easily. At the same time, they can yield upon contact with objects instead of resisting rigidly.

Baseball as a stress test

RAI demonstrates how well this works with various baseball exercises. When throwing, AthenaZero reaches a speed of around 111 kilometers per hour with a tennis ball. With a significantly heavier baseball, it manages around 77 kilometers per hour.

AthenaZero also shows its capabilities when catching and batting. It can catch baseballs at up to 66 kilometers per hour. However, according to RAI, this was only successful reliably at speeds of up to around 48 kilometers per hour. At higher speeds, the success rate dropped significantly.

For the batting attempts, the researchers used tennis balls. In 33 attempts with a ball speed of around 48 kilometers per hour, it hit 27 times. This corresponds to a hit rate of just under 82 percent. According to RAI, even a deviation of two milliseconds in the timing of the swing could lead to a miss.

The catching and batting attempts took place at a distance of only 7.3 meters. In professional baseball, the distance between the pitcher's mound and home plate is about 18.4 meters. The robot has correspondingly less time to react during the shorter throws.

The researchers also had two robots play with each other. They managed up to eight volleys in a row. With a human partner, AthenaZero managed up to twelve. Between catching and throwing again, it took about eleven seconds for each of the two robots. An external infrared camera system tracked the flight paths of the balls during the tests.

AthenaZero learns to juggle in minutes

In another demonstration from May 2026, RAI shows that the robot can also juggle. For this, it uses its multi-jointed hands and camera-based perception.

According to the institute, the robot masters various juggling patterns. These include the classic cascade and patterns where the balls follow different trajectories.

The short learning time is remarkable. AthenaZero is said to be able to learn basic movement sequences like the cascade in less than ten minutes through direct training on the robot. Already learned movements can then be combined for other juggling patterns.

There is still work to be done for the industry

However, the technology has its limits. According to RAI, AthenaZero is less suitable for tasks where it has to hold heavy objects for a long time with an extended arm. In such cases, its motors can overheat. Tasks that require particularly rigid guidance along an exactly defined path are also among its weaknesses.

The institute sees the potential primarily in activities that require fast movements and changing forces. In the future, this could include assembly work where robots must react to unexpected contacts.

Header image: RAI Institute / YouTube

This article has no likes yet.


User Avatar
User Avatar

My interests are varied, I just like to enjoy life. Always on the lookout for news about darts, gaming, films and series.


News + Trends

From the latest iPhone to the return of 80s fashion. The editorial team will help you make sense of it all.

Show all

These articles might also interest you

Show all

These articles might also interest you

Comments

Avatar