Video Friday: Humanoid Learns Tennis Skills Playing Humans
Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar of upcoming robotics events for the next few months. Please send us your events for inclusion. ICRA 2026 : 1–5 June 2026, VIENNA Summer School on Multi-Robot Systems : 29 July–4 August 2026, PRAGUE Enjoy today’s videos! Human athletes demonstrate versatile and highly dynamic tennis skills to successfully conduct competitive rallies with a high-speed tennis ball. However, reproducing such behaviors on humanoid robots is difficult, partially due to the lack of perfect humanoid action data or human kinematic motion data in tennis scenarios as reference. In this work, we propose LATENT, a system that Learns Athletic humanoid TEnnis skills from imperfect human motioN daTa. [ LATENT ] A beautifully designed robot inspired by Strandbeests. [ Cranfield University ] We believe we’re the first robotics company to demonstrate a robot peeling an apple with dual dexterous humanlike hands. This breakthrough closes a key gap in robotics, achieving bimanual, contact-rich manipulation and moving far beyond the limits of simple grippers. Today’s AI models (VLMs) are excellent at perception but struggle with action. Controlling high-degree-of-freedom hands for tasks like this is incredibly complex, and precise finger-level teleoperation is nearly impossible for humans. Our first step was a shared-autonomy system: rather than controlling every finger, the operator triggers prelearned skills like a “rotate apple or tennis ball” primitive via a keyboard press or pedal. This makes scalable data collection and
In a world where technology continues to push the boundaries of what is possible, the latest advancements in robotics are bringing us closer to creating humanoid robots that can perform complex tasks with ease. This week's Video Friday selection from IEEE Spectrum robotics highlights two groundbreaking developments that demonstrate the incredible potential of these machines.
First, the video showcases a humanoid robot that has learned to play tennis using imperfect human motion data. Tennis is a sport that requires a high level of agility, coordination, and precision, making it a challenging task for robots. However, the researchers behind this project have developed a system called LATENT, which stands for "Learns Athletic humanoid TEnnis skills from imperfect human motioN daTa." By leveraging this system, the robot is able to replicate the dynamic movements of human athletes and engage in competitive tennis rallies with a high-speed tennis ball.
The development of LATENT is significant because it addresses a major challenge in robotics: the lack of perfect humanoid action data or human kinematic motion data in tennis scenarios. Previously, robots struggled to reproduce the complex behaviors exhibited by human athletes, limiting their ability to perform tasks that require a high degree of dexterity and adaptability. LATENT's approach of using imperfect human motion data as a reference allows the robot to learn and improve its skills over time, even when the data is not perfect.
In addition to the tennis-playing robot, the video also introduces a robot designed by Cranfield University that can peel an apple using dual dexterous humanlike hands. This breakthrough represents a significant milestone in robotics, as it closes a key gap in the field by achieving bimanual, contact-rich manipulation. Traditional robots often rely on simple grippers, which are limited in their ability to perform complex tasks that require fine motor skills. The new robot, inspired by Strandbeests, goes far beyond these limitations, demonstrating the potential for humanoid robots to perform tasks that were previously thought to be impossible.
The researchers behind this project have developed a shared-autonomy system that allows operators to trigger prelearned skills, such as "rotate apple or tennis ball," via a keyboard press or pedal. This approach simplifies the control process, making it more manageable for humans and enabling scalable data collection and reinforcement learning (RL) training. The system's success is further enhanced by the use of a novel AI model called "MoDE-VLA" (Mixture of Dexterous Experts). This model fuses vision, language, force, and touch data by utilizing a team of specialist "experts," which stabilizes and enhances control in high-dimensional spaces.
The combination of these two innovations—LATENT and the apple-peeling robot—demonstrates the incredible potential of robotics to revolutionize various industries and everyday life. As researchers continue to push the boundaries of what is possible, we can expect to see even more remarkable advancements in the years to come. These developments not only showcase the incredible capabilities of humanoid robots but also highlight the importance of interdisciplinary collaboration and the integration of different fields, such as AI, robotics, and human motion analysis, to achieve groundbreaking results.
In conclusion, this week's Video Friday selection from IEEE Spectrum robotics offers a glimpse into the exciting future of humanoid robotics. With advancements like LATENT and the apple-peeling robot, we are witnessing a new era in which machines are not only capable of performing complex tasks but also learning from imperfect data and adapting to new situations. As these technologies continue to evolve, they hold the promise of transforming industries, enhancing productivity, and improving our daily lives in ways we can only begin to imagine.
