If human eyes are the window to the soul, then what lies inside robotic eyes?
While scientists behind new research in biomimetic eyes may not be interested in the soul of their creation, improvements in robotic eyes could bring them closer than ever to their biological counterparts.
By embedding light-sensitive receptors directly into the surface of a 3D sphere the researchers could design robotic eyes with a sensitivity much closer to that of biological human eyes.
The researchers also gave these robotic eyes an upgrade, making them even quicker to react to light and potentially even better at resolving images than the real deal. Say goodbye to blind spots.
"Mimicking human eyes, artificial vision systems are just as essential in autonomous technologies such as robotics."
In research published Wednesday in the journal Nature, researchers describe why the human eye is such a sought-after target.
"Biological eyes are arguably the most important sensing organ for most of the animals on this planet. In fact, our brains acquire more than 80 percent of information about our surroundings via our eyes," write the authors. "Particularly, the domed shape of the retina has the merit of reducing the complexity of optical systems by directly compensating the aberration from the curved focal plane. Mimicking human eyes, artificial vision systems are just as essential in autonomous technologies such as robotics."
The work was done by scientists at the Hong Kong University of Science and Technology, UC Berkeley, and the Lawrence Berkeley National Laboratory.
Unlike the compound eyes of insects, which use multiple small lenses to process light, the camera-like eyes of humans use a single lens and the rounded shape of the eyeball to focus incoming light into crisp images. This gives us vision that is about 100 times better than insects and is a characteristic that scientists aim to create in robots, or bionic, eyes.
But capturing the essence of human eyes has been a challenge, particularly when it comes to mimicking the round shape of the eyeball. Previous designs have tried to first build flat, light-sensitive materials that they then curve afterward to form the eyeball. But, in order for the material to correctly fold, these previous models had much less densely packed light-sensitive receptors than a true human eye, making them less powerful.
But this, write the authors, is exactly the improvement they made. The team designed an artificial retina by having light-sensitive nanowires built directly in the pores of a hemispherical aluminum oxide membrane. Because this membrane wasn't transformed from a 2D to a 3D shape, the team could more densely pack these sensors than was possible in previous designs.
Between this artificial retina, which made up the "back" of the eyeball, and the symmetrical dome and lens that completed the front half of the eyeball, scientists filled their eye with gel-like ionic liquid that gave power to the nanowires in the retina.
This fluid is analogous to the squishy vitreous humor floating around in our eyes.
When testing the ability of this artificial eyeball the researchers found it was capable of detecting on average of 86 photons per second -- a rate on par with that of a human eye. These artificial eyes were also able to recover from light detection even faster than human eyeballs, taking only 40 seconds to recover while a human eye would take closer to two-and-a-half minutes. The robotic eye was also able to accurately capture, or "see", the letters "E," "Y," "E."
As the researchers continue to refine this proof-of-concept design, they write that future iterations may even have better resolution than human eyes themselves.
One distinct structural difference is that our artificial retina is a front-side illuminated, back-side contacted device so there is no “blind spot” on the retina. In contrast, human retina is a back-side illuminated, front-side contacted organ. Zhiyong Fan, of the Department Electronic and Computer Engineering at The Hong Kong University of Science and Technology and a lead author of the study, tells Inverse that the lack of blind is a "distinct structural difference."
There are more differences. These eyes could see lightwaves beyond the range of humans "One more functional difference is that, human eye can only see optical wavelength range from 400 to 700 nm [nanometers], however our current artificial eye can already respond to 200 nm ~ 800 nm wavelength range. In the future, if we choose to use a narrow bandgap semiconductor as photosensing material to build our artificial retina, then infrared light will be visible to the artificial eye," Fan tells Inverse.
"A human user of the artificial eye will gain night vision capability," Fan says.
This could be especially important for robots, or even as prosthetics. Sharper eyesight could have any number of valuable uses—search and rescue, security, even niche fields like sports analysis.
"Currently in our artificial retina, the nanowire photosensor density is about 6 times higher than that in a human retina," Fan tells Inverse. "This means the artificial eye can achieve higher resolution than a human eye. However, we were limited by the back contact electrode size in current work so we have not demonstrated the full potential in terms of resolution at this moment." Once that problem is addressed, Fan says, "a user of our artificial eye will be able to see smaller objects and further distance."
Abstract: Human eyes possess exceptional image-sensing characteristics such as an extremely wide field of view, high resolution and sensitivity with low aberration. Biomimetic eyes with such characteristics are highly desirable, especially in robotics and visual prostheses. However, the spherical shape and the retina of the biological eye pose an enormous fabrication challenge for biomimetic devices. Here we present an electrochemical eye with a hemispherical retina made of a high-density array of nanowires mimicking the photoreceptors on a human retina. The device design has a high degree of structural similarity to a human eye with the potential to achieve high imaging resolution when individual nanowires are electrically addressed. Additionally, we demonstrate the image-sensing function of our biomimetic device by reconstructing the optical patterns projected onto the device. This work may lead to biomimetic photosensing devices that could find use in a wide spectrum of technological applications.