A ROBOT that can play with a yo-yo in the dark may sound as useful as a chocolate teapot, but it could be an important step towards creating highly mobile, low-cost robots.
Studying the motions involved in rhythmic activities such as playing with a yo-yo and juggling could help make robots more stable when walking, says , a specialist in bioinspired computing at University College London. 鈥淭he cyclic dynamics of the yo-yo may share some properties with the cyclic behaviours of limb movement,鈥� he says. 鈥淪o if we can get robots to play yo-yos more effectively we may be able to get them to walk and run more effectively, too.鈥�
鈥淭he cyclic dynamics of the yo-yo may share some properties with the cyclic action of limb movement鈥�
Advertisement
A team led by roboticist Leon Zlajpah at the Jo啪ef Stefan Institute in Ljubljana, Slovenia, has previously developed software that allowed a robot鈥檚 arm to control a simulated yo-yo on a computer screen. The robot used a camera to watch the progress of the yo-yo, allowing a control system to pull it upwards just before it reached the end of its string.
However, the sophisticated sensors and computer processing needed for a robot to perform tasks using their vision alone are expensive, says , a roboticist at the Israel Institute of Technology in Haifa. Cheaper robots that can work in all conditions, including darkness or bad weather, will need to feel their way through tasks such as walking or running.
So Zacksenhouse and her colleagues have developed a robot that can use a yo-yo without any visual information. Instead it responds to the change of force, or kick, just before the yo-yo reaches the end of its unwinding string.
Their trick, to be reported in IEEE Transactions on Robotics, is to use the sensors on the robot鈥檚 arm to detect this kick and feed it back into the robot鈥檚 control system. It is then used to tune the electronic circuit that drives the arm, so that it locks itself to the motion of the yo-yo. 鈥淲e have learned to stabilise the yo-yo motion using the simple force feedback that arrives once every cycle, instead of continuous, complicated visual feedback,鈥� says Zacksenhouse.
