Engineers from Carnegie Mellon University incollaboration with specialists from the University of Minnesota made a real breakthrough in the development of non-invasive methods for controlling a robotic device. Using a non-invasive neurocomputer interface (BCI), scientists created the world's first robotic arm, controlled by the human brain and with the ability to track the mouse cursor on a computer screen. As noted by Tech Explore portal, which reports on the development, the possibility of using a non-invasive method of controlling robotic devices will have a wide range of applications. For example, this technology can be very useful for paralyzed people.
The disadvantages of modern neurocomputer interfaces
In the past, neurocomputer interfaces have alreadyshowed a very high accuracy of control of robotic objects, but for this purpose special implants were used, which were implanted into the human brain and tracked the necessary signals. The introduction of these implants is a very complicated and rather dangerous task requiring surgical intervention. In addition, such devices are very expensive, but in fact can be very limited in their capabilities. Therefore, such mechanisms are used in very rare cases.
One of the main current development tasksNeurocomputer interfaces are the creation of less invasive, and better completely non-invasive technologies that allow paralyzed people to control their surroundings or robotic limbs with the power of their thoughts, which will help millions of people around the world improve their quality of life.
The problem is that the usenon-invasive neurocomputer interfaces that collect information about brain signals through external electrodes, rather than through implanted into the brain chips, are accompanied by transmitting together the brain signals the amount of "noise", which reduces control accuracy. Therefore, non-invasive methods of mental control of robotic devices are losing heavily to technologies using implants. Despite this, developers are not giving up and are trying to create more accurate management methods that will not require surgery.
And it seems that the head of the department of biomedical engineering at Carnegie Mellon University, Professor Bin Heh succeeded.
“In the development of robotic devices withconsiderable success has been achieved by controlling the power of thought through special brain implants. It really is. However, the key goal of such developments is the creation of non-invasive methods. The advances in neural decoding and the practical utility of non-invasive control with robotic arms will have a crucial impact on the development of non-invasive neurobiotics, ”he comments.
The world's first power of thought robotic arm
Using new methods of sensory technology andmachine learning Xe, along with his team, was able to gain access to the deep brain signals, achieving high precision control of roboruki. Thanks to non-invasive neuroimaging and a new paradigm of continuous pursuit, the system learned to overcome noisy EEG signals, which led to a significant improvement in neural decoding and facilitated the continuous control of robotic devices in real time.
Using non-invasive BCI to controlwith a robotic hand that tracks the cursor on a computer screen, Xe with the team for the first time demonstrated that the manipulator can now continuously follow the cursor. Previously, such devices followed the movement of the cursor with sharp, discrete movements, as if they were trying to “catch up” with commands from the brain — now the limb follows the cursor smoothly and continuously.
In an article published in the journal ScienceRobotics, the team describes the creation of a new platform that aims to improve the brain and computer components of BCI by increasing user involvement and learning, as well as improving the spatial resolution of non-invasive neural data by visualizing EEG sources.
The article states that the team’s unique approach toSolving this problem has improved the BCI training system by almost 60 percent for traditional centering tasks, and also improved the quality of continuous tracking of the computer cursor by more than 500 percent.
To date, the system has been tested withthe participation of 68 able-bodied people (up to 10 sessions for each), people (up to 10 sessions for each subject), including virtual device control and robotic arm control for continuous pursuit. Clinical trials of technology with the participation of these patients, scientists plan to start conducting in the very near future.
Video of one of the experiment participants,using a virtual cursor using a neurocomputer interface, as well as a robotic arm, controlled by the power of thought to perform the task of continuously monitoring a randomly moving target
“Despite technical problems withUsing non-invasive signals, we are fully committed to bringing this safe and cost-effective technology to people who can benefit from it, ”says Heh.
“This work is an important step inthe development of non-invasive computer-brain interfaces - a technology that can someday become a universal assistive system that helps people like smartphones. ”
You can discuss the news in our Telegram-chat.