Researchers in Switzerland, led by a Greek dispersal engineer specializing in computers and robots, have taken an important step in creating robots that can only be controlled by the human mind. This, among other things, is intended to be used by quadriplegic patients “imprisoned” inside their body and unable to speak or make the slightest movement.
People e.g. who have suffered severe spinal cord injuries, often have permanent neurological damage and severe motor disabilities, so that they can not even hold an object. Scientists in many countries have been working for years to develop systems that will help these patients perform certain tasks. One solution is for robots to be able to help these patients by doing certain things on behalf of paralyzed people. But first a way must be found for robots to take human commands in real time and respond to them correctly.
Researchers from the Laboratory of Algorithms and Learning Systems of the Federal Polytechnic School of Lausanne (EPFL), led by the Greek Jason Batzianoulis, who published the relevant publication in the journal “Communications Biology”, have developed a computer robots “mentally”, using electrical signals sent by the patient’s brain. The patient does not need to say anything to the robot or press a button, as he can only move the robot with his mind and get it to do some tasks.
The researchers first programmed a robotic arm to avoid obstacles and then gradually improved its mechanism to become more precise, developing an algorithm that can adjust the robot’s movements. To achieve this, the algorithm is connected to a special helmet with electrodes that is worn on the patient’s head and “reads” through electroencephalogram the brain activity of the latter. All the paralyzed patient does is look at the robot and think about what he would like the robot to do instead of him. If the robot makes a wrong move, the patient’s brain will send an error message as if the patient were saying “no”. The robot will then realize that it is doing something wrong, but the point is to understand what is wrong with it (eg that it is too close to an object or that it is still too far away).
In order for the robot to find the correct answer, the error message from the patient’s brain is properly “translated” by the algorithm through artificial intelligence (machine learning) and fed to the robot. The whole process is done very quickly, until the robot “realizes” what exactly is the human desire, so that he can adjust his behavior accordingly. The development of error signal detection technology has been a major challenge for researchers.
According to Batzianoulis, who is a graduate of the Department of Electrical and Computer Engineering of the Aristotle University of Thessaloniki (AUTh 2011) with a PhD in Robotics and Neuroprosthetics from EPFL (2019) and from 2021 an engineer in the Swiss company Ansys which was particularly difficult in our study, was to make the connection between the patient’s brain activity and the robot control system. In other words, to “translate” the patient’s brain signals into actions performed by the robot. We achieved this by using machine learning to associate a specific brain signal with a specific task. We then correlated the tasks with individual robot control commands, so that the robot can finally make what the patient has in mind.
The researchers hope that eventually their algorithm, among other things, will be used to control the paralyzed with their “smart” wheelchairs, something for which, however, several technological difficulties still need to be overcome.
Link to the scientific publication:
https://www.nature.com/articles/s42003-021-02891-8
Source: RES
© 2020 DERA ATHENS 9.84
Switzerland: One step closer to robots controlled by the thought of a paralyzed man – Αθήνα 9.84
