Share this | | | | -A A A+
8+
Gallery 7

Robot Hockey

a yellow robot arm with a hocky puck attached at the end of it's arm sitting on a blue air hockey table
A blue hockey table with a tv screen above it and perspex glass enclosing the a robot arm at the back.

Robotic technology allows a mechanical arm to detect and respond to human actions in a challenging game of air hockey.

How it works

Compete against a robotic arm in a game of air hockey. A screen displays the number of points (goals) won by the humans or the robot during a game.

Things to try or ask around the exhibit

  • How does the robot 'see' the puck and know where to position its arm?
  • How can the computer tell the difference between the puck and your paddle?
  • How does the robotic arm physically move across the goal?
  • Observe the lighting, the sensors and work out how the robot computes the movement of the puck.

Background

This exhibit demonstrates how sensors detect the position and speed of the puck and how the robot computer's program logic instructs the robotic arm to position itself and block a goal, then hit and return the puck back to your end.

Retroreflective material on the puck reflects light straight back to the source. The high speed camera mounted above the table is surrounded with LEDs that shine light onto the table, so light is reflected off the puck and straight back to the camera. This happens because of the way the material is constructed: minute surfaces are joined at right-angles (90 degrees) to each other, so any incidental light is reflected back out at the same incidental angle (called the Law of Reflection). Objects like your hand or the table are not constructed with the necessary minute right-angle surfaces, so the light incident on them is not reflected straight back at the source. To the robot’s computer, the puck appears much, much brighter than anything else.

Both the puck and the paddle have retroreflective material. The paddle has a smaller reflective area than the puck, and the computer recognises that a smaller bright spot is the position of the paddle, rather than the puck. As the computer only ‘sees’ in pixels, it is able to calculate the number of pixels in a bright patch.

A system called 'Binary Filtering' is then used where any pixels below a certain level of brightness are rejected from the computer’s imaging process. After this process, the pixels that remain behind are the puck and so the coordinates of the puck can be determined and a return shot calculated!

The high speed camera captures around 167 times per second, giving the computer a highly detailed assessment of the position of the puck, and allowing the computer to calculate the puck’s angles of incidence when it hits and rebounds from a surface, the speed of the puck and where the robot should position its arm to block and return the puck back to the visitor.

Finding the science in your world

Our eyes detect light (just like the exhibit’s high speed camera’s sensors) and send a message to our brain. Our brain translates this information and judges the best way for our muscles to respond. This includes the angle the paddle is held, the movement of our arms or body, as well as the amount of force applied to the action. All these calculations are made very quickly in our brain and feel instinctive to us. The computer has to do all these calculations too, and then pass the information to the robotic arm.

The retro reflective material on the puck is the same as that used on life jackets and safety vests and traffic signs, because it reflects light straight back to the source, making things more visible. You can see the traffic signals because your car’s headlights are shining on them.

Robotics differ from animatronics. In robotics, sensors detect light, heat, movement and colour which is then interpreted by the robot's computer to decide what action needs to be taken and respond (a bit similar to humans responding to a situation). Most technology in automated factories tends to be based on robotics. In animatronics, the 'robot' performs pre-programmed actions that do not depend on detecting and responding to stimuli. Most 'robots' in theme parks and toy shops are based on animatronics.