Tropes

R_news.discovery.com 2015 02309.txt.txt

#Robotic Limb System Learns From Its Mistakes The science of brain-machine interface, or BMI, has made enormous leaps in the last few decades. For patients with significant motor impairments, BMI tech allows the use of artificial limbs by way of electrodes connected to the brain. With training, patients can move and control their arms or legs again, literally by thinking about it. It life-changing technology, to be sure, but the training involved is arduous, inefficient and it doesn work for everyone. New research coming out of Europe may change all that, ushering in a new generation of BMI systems. In a study published today in Nature Scientific Reports researcher Jose Millán of the Center for Neuroprosthetics at EPFL unveiled a new kind of artificially intelligent BMI system that learns from its mistakes. It works like this: Existing neuroprostheses require the user to generate specific brainwave activities for particular motions xtend left arm, for instance. The brain activity is picked then up through an electroencephalogram and translated into instructions for the prosthetic limb. If the triggering brain activity isn precisely correct, the desired action fails and the brain emits an electrical signal signifying the failure. Millán team has found a way to make use of those error signals by teaching the machine itself to learn from mistakes. When the neuroprosthetic system detects the error message from the brain, it understands that the action was unsuccessful and adjusts movements accordingly. If a patient is trying to grasp an object for example, the intelligent prosthesis will make adjustments and increase precision on its own until no error messages are generated and the goal is achieved. he paradigm shift lies in the use of these signals to relieve the subject from the tedious task of learning, according to the press materials issued by EPFL. his new approach could be the source of a new generation of intelligent prostheses, able to learn a wide range of movements. Indeed it is theoretically possible to learn and master quickly enough a multitude of motor movements, even the most complex ones. h

< Back - Next >