
Human thoughts could come to help robots fix their mistakes and actions.
STATES CHRONICLE – A team of scientists developed a technology which allows robots to fix their mistakes. How? With help from the human thoughts transmitted to them.
Presently, scientists are developing what can one day become a staple technology. They are working on a system through which humans can start controlling robots with just their minds.
Research on the matter is being carried out at the CSAIL. This is MIT’s Computer Science and Artificial Intelligence Laboratory. This study also involves Boston University researchers. The team released a research paper earlier this week. They also published a video showcasing the technology.
The researchers are looking to create a system capable of commanding robots simply through human thoughts. Basically, the respective people could be able to control them with just their minds.
Presently, the team developed a feedback system. As it is, this currently enables connected robots to fix their mistakes. The system revolves around an EEG or electroencephalography monitor. This displays brain activity data.
The system can scan for a set of specific brain signals. These are called ErrPs or “error-related potentials”. These usually appear when the brain detects a mistake. This new technology enables a real-time feedback.
One which is made possible by machine-learning algorithms. These make the system capable of reading and classifying brain waves. Such human thoughts can be classified in just around 10 to 20 milliseconds. And connected robots will be able to pick up the signals. Based on them, they will adjust their behavior and fix potential mistakes.
Daniela Rus, the CSAIL Director, went to offer details. She is also the senior author of the research paper. According to her, the system works as follows. A person is connected to the system. Then, all they have to do is watch a robot. Based on its actions, the person only has to mentally agree or disagree with its activities.
Rus also points out the following. The involved participants do not require training. These human thoughts can come as they want. They do not have to adapt to the machine. Instead, the robot will have to adapt to the human thoughts.
More details on the current study will probably emerge in May. The study paper was accepted and will be presented during the ICRA. This is the IEEE International Conference on Robotics and Automation, which will be taking place in about two months.
As it is, the current study is still in the early stages of the full human brain – robot control relation. Monitoring ErrPs or other brain waves is just one step. A more advanced operation than just reading brain signals would be needed.
Nonetheless, this research does open up more than just one path. It could also help develop a more intuitive human-robot relation and interaction. One which could have application in the real world. For example, in autonomous driving systems or in managing factory assembly line robots.
Image Souurce: Wikimedia