Humanoid Robots can feel now
With the development of smart technologies, new types of Humanoid Robots are having the ability to feel materials . Even this types of Humanoid robots can be more touch sensitive then humans . This types Of Humanoid Robots can feel with the help of sensors, software and actuators. This smart technologies have the abilities to identify different materials by touching them. Already IBM researchers predicted that we will be able to touch real object faraway from us by our mobile phone or computer screen within five years(Smart future technology – virtual touch of real objects ).This invention is a step to that prediction.
Researchers of the University of Southern California’s Viterbi School of Engineering published a article in Frontiers in Neurorobotics related to this types of Humanoid robots .In this article they wrote that how specially designed Humanoid Robots can surpass humans in identifying natural materials by their texture.
This sensitive types Humanoid robots have special types of sensors in their hand. This are tactile sensors built to act as human fingertip. This tactile sensors are Biotechnology sensors with produce vibration when it explore a material like human fingertips. This types of sensors have a soft, flexible skin. This skin is over a liquid filling. There is even fingerprints on the surface of the skin. This fingerprints enhance the sensitivity of sensors to vibration. When the finger of Humanoid Robots slides over any material, the skin starts to vibrate. This vibration I detected by a hydrophone. This hydrophone is in the finger of Humanoid robots , as like as the bone of the finger. Our human fingers use similar types of pattern to identify materials, but this types Humanoid robots are more sensitive.
But to identify any material, still required an algorithm. This types Humanoid robots use a newly designed algorithm to explore a material. This algorithm is based on human strategies. While exploring a material humans concentrate on it’s traction, roughness, and fineness. Humans also take informative decisions based on the sequence of the exploratory movements. This exploratory movements find out the most information for the task. They depends on the past knowledge about those exploratory movements. This way they guess what a material may be. According to a famous theorem by Thomas Bayes ,decisions might be made from the information obtained during these exploratory movements. This exploratory movements done by human based on their prior experiences. But until now there was no way to decide robots have to made which exploratory movement to identify a material. Gerald Loeb Professor of Biomedical Engineering and recently Jeremy Fishel a graduated doctoral student, the writers of that article describes their new theorem for solving this general problem as “Bayesian Exploration.”
Their specialized humanoid robot was trained to identify 117 different types of material such as fabrics, stationary, and hardware stones. The Humanoid robot has been able to correctly identify the material 95% of times , when it was given materials at random. It was only rarely confused by pairs of similar textures what human could not distinguish at all by their exploratory movements.
This types of humanoid robots can be used in companies to identify different materials.
Content source: http://www.sciencedaily.com/releases/2012/06/120618194952.htm