Toyota’s Basketball Robot CUE7 Makes Headlines
When you think of Toyota, it’s usually about their cars like the Camry or the Tacoma. But picture this: a 7-foot-2 robot casually shooting free throws during a basketball game. That’s exactly what went down at Toyota Arena in Tokyo, where around 8,400 fans witnessed the event live.
This robot, named CUE7, amazed the audience as it effortlessly rose, dribbled, and sank free throws on its own. The crowd erupted in applause, likely leaving Toyota’s engineers holding their breath as they watched their creation debut in such a thrilling setting.
So, you might wonder, why would a car manufacturer be interested in building a basketball robot? And what’s the connection to you? It’s more significant than you might think.
CUE7 Changes the Game
What really differentiates CUE7 from earlier models is its clean slate approach. According to Tomohiro Nomi, who leads humanoid robot research at Toyota, they decided to start anew, discarding previous designs entirely. This isn’t just a minor detail; it’s substantial.
The CUE series kicked off back in 2017 with passionate Toyota employees, and it eventually turned into an official research initiative. Over the years, the team built impressive tech milestones, including CUE3’s record of 2,020 consecutive free throws in 2019. CUE6 even set records for the longest basketball shot made by a robot.
While that legacy is noteworthy, the real shift with CUE7 lies in how it learns.
AI Learns to Shoot
Previous models used a system called model predictive control, where humans meticulously programmed every move. Although it worked well and even earned records, it limited growth since every alteration required fresh coding.
CUE7, however, operates on reinforcement learning. Basically, it learns to shoot by trial and error, adjusting based on experiences rather than relying solely on pre-set instructions. It’s almost like comparing a player who knows every play to one who reads the game live; CUE7 is adapting in real-time.
What’s Inside CUE7?
Standing 7 feet 2 inches tall and weighing about 163 pounds—40% lighter than its predecessor—CUE7 boasts a streamlined design with fewer axles and only two wheels for smoother movement. Its standout moment? Rising gracefully from a seated position, which impressed the audience immensely.
For sensing and aiming, the robot has lidar sensors that help it gauge its surroundings and stereo cameras to determine distances. A high-performance battery, drawing on Toyota’s racing background, powers it.
Here’s the fascinating part: the robot calculates the hoop’s distance, figures out angles, and shoots while adjusting its force according to previous attempts.
Learning Through Real Performance
The secret sauce behind these lifelike movements? Using data from human actions to train CUE7, making it appear less robotic. What’s more, as it handles dribbling and shooting simultaneously, its actions are both fluid and precise.
Testing robots in environments like a basketball stadium is crucial, according to Tomohiro Nomi. He sees it as an invaluable opportunity to create excitement and joy for people.
Why This Matters to You
While you’re unlikely to buy a robotic basketball player anytime soon, the AI driving CUE7 could significantly influence manufacturing and robotics. Basketball stresses various capabilities that manufacturing robots have more trouble with, such as targeting and coordination. It serves as a practical test for all these skills in a clear-cut environment.
The technology could eventually be adapted for factory robots that adjust to shifting production needs or vehicles that react more smoothly to unpredictable road conditions. In essence, Toyota views CUE7 as more than just a fun demo; it’s a foundational step towards broader applications in real-life scenarios.
AI Learning: A Broader Impact
When Toyota teaches a robot to shoot hoops, they aren’t merely playing games; they’re laying the groundwork for machines capable of learning physical skills in uncertain settings. The implications here stretch far beyond basketball courts.
So, if this robot can learn to outshoot many humans through pure trial and error, it begs an intriguing question: Are there still physical skills you believe a machine could never grasp?
