Improbable takes a look at the RO-MAN Workshop 2012 – the 21st IEEE International Symposium on Robot and Human Interactive Communication, which was held September 9-13, Paris, France.
Part 1 : Avoiding a poke in the eye
NOTE: No students were harmed during this experiment – conducted at the Academy for Safety Intelligence, Nagoya University, Japan.
“The participant wearing protective glasses and the robot were separated by a working table. To minimize harming the participant, the original end- effectors for picking up and placing mechanical parts were replaced with square pyramid-shaped flexible polyurethane foam.”
And then :
“… a situation is created in which the sharp end effector tip of a robot suddenly approaches the eyes of a facing participant.”
But would the experimental participant move out of the way in time? That might depend, to a large extent, on their Reaction Time (RT).
“The results showed that there are individual differences in the avoidance RT, maximum avoidance acceleration, and maximum avoidance speed. The results also showed that the avoidance RT is negatively correlated with the maximum avoidance acceleration and maximum avoidance speed, while that the latter two have a strong positive correlation.”
Thus, explain the researchers :
“… further investigations into avoidance action characteristics should be pursued with a focus on people who have a lower ability to avoid harm.”
UPDATE : Needless to say, the experiment could not be allowed to go so far as to actually poke a student in the eye, even with a foam-encased gripper – so the research team set about another study (following-up the RO-MAN paper) in which a robot’s metallic end-effector was fully exposed. And this time the robot jabbed it directly at an unprotected eyeball, causing considerable damage. The eyeball did not belong to a student, but rather, a recently-deceased pig.
“Specifically, the eyelid served as a safeguard. In regard to the collision position, the severity of injury to the sclera was estimated to be lower than that to the cornea. In regard to the collision angle, the severity decreased with an increase in the angle from the normal direction of the eyeball. This was a very interesting result.”
See: Basic Experiments on Collision of Sharp Mechanical Hazards against Eye for Estimation of Injury Severity, (Proceedings of the 2012 IEEE International Conference on Robotics and Biomimetics, pp. 1912-1917, Guangzhou, China, 2012, Dec. )
COMING SOON : RO-MAN’s dancin’ ‘bots