TECHNOLOGY TECHNOLOGY Technologies of FORPHEUS

01Updated points for 2017

1 1

POINT 1The ability to serve.

OMRON has succeeded in adding a new toss function to the robot. It can now assist during service, something humans often have trouble with. Through this assistance the robot can help built trust and unity with its opponents.

2 2

POINT 2The ability to predict a smash.

It is now possible for the robot's AI to predict, from the movements of its opponent, the type of shot they plan to play. The ability to predict when a smash is coming allows a more realistic playing experience, especially during rallies.

3 3

POINT 3The ability to produce sounds

The robot is now capable of producing various sounds appropriate to the state of the match. This added sensory experience makes it much more enjoyable and immersive for the user.

TECHNOLOGY TECHNOLOGY Technologies of FORPHEUS

02The core technologies of Forpheus.

Image detection

A three-camera system identifies the ball and evaluates players.

Cameras mounted on the left and right of the robot allow it to identify ping-pong balls in a 3D space in an identical way to the human visual system. The ball's location is detected up to 80 times per second. In addition to the two side cameras, Forpheus has a third central camera which is used to evaluate the player and, based on their movements, judge their ability level.

Arm control

A high speed robotic arm which moves in response to its AI controller.

Forpheus utilizes a repurposed arm from an industrial robot used in product packaging.
The arm is controlled by the AI through a 5-axis motor system to swing the paddle at the ball. Paddle grip, ball hit location and arm position are all controlled to within 0.1mm.

Trajectory prediction

Predicting where the ball will go.

Forpheus calculates the speed and rotation of the ball up to 80 times per second allowing it to predict the ball's trajectory with high accuracy. Forpheus can calculate the trajectory of the ball the moment its opponent hits it (red trajectory). The robot preemptively then moves its paddle to a position where the ball will be when it needs to return it. Forpheus also calculates the trajectory of the balls it will be hitting back (blue trajectory) and adjusts the difficulty of its return shot based on the ability of the opponent.

Motion controller

The brain of Forpheus which tells it how to hit the ball.

The motion controller at the heart of Forpheus must be able to make decisions on timing and direction within a thousandth of a second. The directional error rate for the return balls is within ± 10 cm. The controller used in Forpheus is an OMRON NJ series which is typically used in production facilities. All user interface technologies such as the touch panels for starting and switching modes and the servo controllers for the motors also use OMRON's latest technology.

Machine learning by AI (deep learning)

Improving the system's ability to judge the level of its opponent.

By evaluating the opponent's movements and the motion of the ball the AI can judge whether the opponent is experienced or a beginner. Forpheus will dynamically adjust its ability level based on its opponent to try and maximize the potential for a successful rally.

PLAY