"One of the challenges with robotic fingers is ensuring that they can be controlled precisely enough to softly land on biological tissue," said Hangue Park, assistant professor in the Department of Electrical and Computer Engineering. "With our design, surgeons will be able to get an intuitive sense of how far their robotic fingers are from contact, information they can then use to touch fragile structures with just the right amount of force."
Robot-assisted surgical systems, also known as telerobotic surgical systems, are physical extensions of a surgeon. By controlling robotic fingers with movements of their own fingers, surgeons can perform intricate procedures remotely, thus expanding the number of patients that they can provide medical attention.
To move their robotic fingers precisely, surgeons rely on live streaming of visual information from cameras fitted on telerobotic arms. Thus, they look into monitors to match their finger movements with those of the telerobotic fingers. In this way, they know where their robotic fingers are in space and how close these fingers are to each other.
However, Park noted that just visual information is not enough to guide fine finger movements, which is critical when the fingers are in the close vicinity of the brain or other delicate tissue.
"Surgeons can only know how far apart their actual fingers are from each other indirectly, that is, by looking at where their robotic fingers are relative to each other on a monitor," Park said. "This roundabout view diminishes their sense of how far apart their actual fingers are from each other, which then affects how they control their robotic fingers."
To address this problem, Park and his team came up with an alternate way to deliver distance information that is independent of visual feedback. By passing different frequencies of electrical currents onto fingertips via gloves fitted with stimulation probes, the researchers were able to train users to associate the frequency of current pulses with distance, that is, increasing current frequencies indicated the closing distance from a test object. They then compared if users receiving current stimulation along with visual information about closing distance on their monitors did better at estimating proximity than those who received visual information alone.
Park and his team also tailored their technology according to the user's sensitivity to electrical current frequencies. In other words, if a user was sensitive to a wider range of current frequencies, the distance information was delivered with smaller steps of increasing currents to maximize the accuracy of proximity estimation.
The researchers found that users receiving electrical pulses were more aware of the proximity to underlying surfaces and could lower their force of contact by around 70 percent, performing much better than the other group. Overall, they observed that proximity information delivered through mild electric pulses was about three times more effective than the visual information alone.
COMPAMED-tradefair.com; Source: Texas A&M University
