Robots that can hit a golf ball on a golf course are not really new. However, it takes a much more complex process for robots to have the skills to play a short game of golf. Researchers at the University of Paderborn in Germany did just that with Golfi, the golf robot that uses neural networks to figure out how to perfect a hit and how hard to hit the ball to get it into the hole from anywhere on the course.Golfi takes a snapshot of the green with the Microsoft Kinect 3D camera and simulates thousands of random shots taken from different locations. The stroke is performed by calculating factors such as the rolling resistance of the grass, the weight of the ball and the starting speed. Paderborn PhD student Annika Junker, on the other hand, told IEEE Research that it only takes five minutes to train Golfi on simulated golf throws.
Better than most golfers
Golfi has a powered arm to hit the ball, but the robot can’t get the ball into the hole every time. Researchers say the robot’s shots are 60-70 percent successful. Although this rate is more successful than most amateur golfers, it should not be forgotten that Golfi has a very short training period.However, the researchers did not set out to build a robot that could compete with the PGA Tour pros. They hope the techniques they used at Golfi could be used in other robotic applications.
26929
