Brainput Project Takes a Load Off Humans' Minds
A recent project conducted by researchers at MIT and other institutions demonstrated a method for computers to help people who are weighed down by intense multitasking. Sensors measured test subjects' brain activity as they juggled multiple tasks. When the machine sensed the person was focusing on one task to the exclusion of the other, it "took over" the neglected task.
05/15/12 3:35 PM PT
A group of researchers from several universities led by MIT have shown that robots controlled mentally by suitably equipped humans who are multitasking can take over some of the workload when needed.
The Brainput project had researchers use a technique called "functional near-infrared" (fNIR) imaging to measure the activity of brains in test subjects while the subjects were keeping track of two robots and trying to prevent them from crashing into walls.
When the robots sensed the driver was fully occupied, their behavior became more autonomous, meaning they made some of the navigational decisions themselves.
Brainput is one of a series of experiments in using the mind to control robots or get devices to perform work that have been going on since the 2000s.
"There's a huge amount of work being done to better blend people and machines, and this is one of those efforts," Rob Enderle, principal analyst at the Enderle Group, told TechNewsWorld. "The idea of a better human-machine interface has been floating around for some time and we're getting close."
The Brainput Experiment
The transmission and absorption of NIR light in human body tissues contains information about changes in the concentration of hemoglobin in those tissues. When a specific area of the brain is activated, the blood volume in that area increases quickly.
The fNIR method measures the level of neuronic activity on the brain by looking at the relationship between metabolic activity and the level of oxygen in a subject's blood.
The Brainput researchers had test subjects remotely navigate two robots to find the best location to transmit data they had collected back to the control center. The robots measured and reported the signal strength in their current location at the navigator's request in order to do this. The participants were told to avoid collisions with obstacles and walls and were advised not to leave either robot idle. This forced them to work with both robots simultaneously.
The researchers found that the robots would take on some of the workload whenever they detected that the test subjects were tending to multiple tasks. When the navigator stopped multitasking, the robots would go back to requiring instructions as to what to do next.
The results led the researchers to conclude that Brainput provides measurable benefits to users with little additional effort required from them.
Real-world applications for this technology could include military operations where unmanned drones are used as well as search-and-rescue operations in which robots are used to explore areas unsafe for humans.
"We're still on the very cutting edge of this stuff," Enderle said. "In the next few years, we're going to see amazing advances."