Andrew Sun-Zhuang
7th Grade, The Evergreen School
Shoreline, WA
VisionStrike: Portable AI-Powered Baseball Pitching Machine
To improve batting ability, baseball players need to face off against a pitcher that can vary its throws. Andrew trained an AI model on 200 images, teaching it to recognize six key points on a batter. His pitching machine ìthrewĂ® strikes and balls with 95 percent overall accuracy for fastball pitching.â€
View Poster
Project Background
Andrew plays baseball, and he’s constantly trying to improve his batting performance. But he’s been less than impressed with pitching machines. “Because they cannot vary pitch location to simulate real-game conditions, batters can anticipate pitch placement and achieve unrealistically high contact rates,” he explains. His hit rate is 95 percent against a batting machine, but only 35 percent in real life. And most batters, Andrew included, can’t practice against a real pitcher every day. He decided to train an AI to produce a better pitching machine.
Tactics and Results
Andrew first built his own pitching machine, using a microcontroller, a Raspberry Pi, a camera, servos and motors. He then trained an AI model on 200 different pictures of a batter—himself. Andrew had the model detect six key points on every image: The end of the bat, the wrist, elbow, shoulder, hip and knee. The model used these points to determine where the bat strike zone should be. When loaded into the Raspberry Pi, the model calculates the right angle once a batter has stood still for two seconds. Then it launches the ball.  Using Andrew’s AI model, the machine pitched with 86 percent accuracy within a strike zone for front-toss pitching. For fastball pitching, it “threw” strikes with 94 percent accuracy and balls with 97 percent accuracy. Andrew’s final prototype cost only $260, a major improvement over pitching machines which usually cost much more.
Beyond the Project
Obviously, Andrew likes baseball, and started in Little League when he was five. Last summer, he played both All-Star Majors baseball and water polo. “I love baseball because of the connections I build with my teammates and the challenges I face during games,” he says. “The pressure of competition pushes me to stay focused, build resilience, and constantly strive to become a better player.” Andrew also enjoys robotics and has participated in several robotics programs. “I have a deep passion for robotics because it is an active, hands-on blend of STEM that can solve real-world problems,” he says.
