Virtual retinal display

A virtual retinal display (VRD), also called a retinal scan display (RSD), retinal projector or retinal projection display, is a display technology that scans or projects a modulated beam of light directly onto the retina of the viewer's eye to form an image, instead of forming the image on a panel or screen that the eye then looks at. Because the light is delivered through the eye's own optics and the picture is painted point by point on the retina, the viewer perceives what looks like a conventional image floating in front of them, even though there is no physical screen. The VRD is a form of Near-eye display and is normally built into a Head-mounted display or a pair of glasses, where its compact optics and see-through potential make it of particular interest for augmented reality.

How it works

A virtual retinal display builds an image the same way an old cathode-ray-tube television does, by sweeping a single spot of light across the field in a raster pattern, but the surface it writes on is the retina rather than a phosphor screen. A typical system has a light source, a way to modulate its brightness, horizontal and vertical scanners, and delivery optics that route the beam into the eye.

The light source is usually one or more lasers or laser diodes, though light-emitting diodes (LEDs) can be used instead. For a full-color display the system combines a red, a green and a blue source so that each pixel can be set to any color. The intensity of the combined beam is modulated pixel by pixel to carry the picture. The modulated beam is then steered in two dimensions, in the original University of Washington work by a fast mechanical resonant scanner for the horizontal sweep and a slower galvanometer mirror for the vertical sweep, so that the spot traces out a complete raster many times per second.

The defining step is the delivery optics. Rather than spreading the light over a panel, the optics bring the scanned beam to a focus at a small point on the eye's pupil. The eye's own cornea and lens then take over and focus the converging light onto the retina, where the scanning paints the image directly on the photoreceptors. Concentrating all of the image-forming light through a single point of the pupil is the classic Maxwellian view arrangement, named after the nineteenth-century physicist James Clerk Maxwell, who observed that light brought to a focus at the pupil produces a bright patch on the retina largely independent of the eye's focus. Because the beam never dwells on one spot but sweeps continuously across the retina at a high rate, the instantaneous energy at any point stays very low.

History

The idea of writing an image straight onto the retina predates the modern VRD. A retinal scanning concept attributed to Kazuo Yoshinaka of the Nippon Electric Company (NEC) appears in a Japanese patent filing from 1986, although it did not lead to a commercial product.

The technology most often meant by the term "virtual retinal display" was developed in the early 1990s at the Human Interface Technology Lab (HIT Lab) at the University of Washington in Seattle, under the direction of Thomas A. Furness III, a researcher who had earlier worked on military helmet-mounted displays. The core invention is recorded in United States patent 5,467,104, titled "Virtual retinal display," which names Thomas A. Furness III and Joel S. Kollin as inventors and is assigned to the University of Washington. The patent was filed on 22 October 1992 and granted on 14 November 1995. The stated goal of the project was a full-color, wide field of view, high-resolution, high-brightness display that was also small, light and inexpensive.

Much of the early laboratory work was funded through a multi-year research contract with Microvision, a Seattle company founded in 1993 that took the exclusive license to commercialize the VRD; some of its early development was financed by United States government defense contracts. In 1998 the HIT Lab's virtual retinal display received the Discover Magazine Award for Technological Innovation in the "sight" category, with the citation naming Furness as a key innovator; the award was presented that July at the EPCOT Center in Florida.

Microvision later turned the technology into shipping hardware. Its first commercial product, the Nomad, was a monocular, see-through, head-worn display that drew a monochrome red image onto the wearer's retina. A version aimed at professionals, the Nomad Expert Technician System, was offered around 2004 at roughly 4,000 US dollars and was marketed to mechanics and other technicians who needed hands-free access to schematics and diagnostic data. Over time Microvision shifted its emphasis from full retinal displays toward laser beam scanning projection engines for pico-projectors and, later, automotive sensing.

Claimed advantages

Supporters of the virtual retinal display point to several benefits that follow from delivering light through the pupil rather than from a panel.

Advantage	Basis
High brightness and color	Laser sources are very bright and have narrow, pure spectral lines, so a VRD can in principle reach high luminance and a wide color gamut, which helps for outdoor and see-through use
Large depth of field	In a Maxwellian view the image is focused through a single point of the pupil, so the picture stays sharp almost regardless of where the eye accommodates, and even for users with short-sightedness, long-sightedness or astigmatism
Compact, lightweight optics	There is no display panel or large magnifier to carry; the image-forming hardware can be a small light engine and scanner, which suits slim glasses and helmet-mounted units
Good for see-through AR	Only a thin beam needs to reach the eye, so most of the field of view can stay transparent, letting synthetic content overlay the real world for augmented reality
Low light power	Because the spot sweeps the retina continuously, only a tiny fraction of a watt reaches the eye, on the order of a milliwatt or less, which keeps power draw low

Because the focus is set by the geometry of the beam rather than by the eye's lens, the VRD largely sidesteps the focus problem that shapes much of Near-eye display design, where a panel sits closer than the eye can comfortably focus.

Challenges

The same physics that gives the virtual retinal display its strengths also creates its main difficulties.

The most fundamental limitation is a small eye box. Since every image point passes through one small point at the pupil, the eye must stay aligned with that point; if it rotates or shifts so the convergence point falls outside the pupil, the image dims or disappears entirely. This narrow tolerance for eye and headset movement is the chief reason retinal projection has remained largely in research and niche products, and much later work has gone into widening the eye box with techniques such as multiple viewpoints or holographic combiners. Closely related is the need for precise optical alignment between the device and the wearer's eye, which makes fit, calibration and slippage more demanding than for panel-based displays.

A further obstacle is the perception of laser safety. Shining a laser into the eye sounds hazardous, and that impression can deter users even when the device is engineered to be safe. In practice the light is kept to very low power and is constantly in motion, so well-designed VRDs are built to meet Class 1 laser standards under IEC 60825, the level treated as safe under all normal conditions of use. Laser safety analyses of retinal scanning displays have calculated the maximum permissible exposure for both normal operation and credible failure modes and found measured power well within safe limits. A practical drawback noted in early systems was a restricted image area and, in monochrome units, the lack of full color.

Modern relevance

Although full virtual retinal displays remain uncommon, the underlying technique of scanning a laser beam to form an image, often called laser beam scanning (LBS), has become a mainstream approach for compact projectors and some AR eyewear. LBS engines use a tiny MEMS mirror to sweep red, green and blue laser light into a raster, and their small size, low power and vivid color make them attractive for pico-projectors and for augmented reality glasses that paint an image on or near the eye.

A widely noticed example was Intel's Vaunt prototype, shown in 2018, a pair of ordinary-looking glasses that used a low-power vertical-cavity surface-emitting laser to project a red image off a holographic reflector on the lens and into the eye. Intel later closed its wearables group, and in December 2018 the startup North acquired the Vaunt patents and technology; North was in turn acquired by Google in 2020.

The closest modern descendants of the true VRD are the retinal scanning eyewear from the Japanese company QD Laser. Its RETISSA Display, launched in 2018, and the later RETISSA Display II project a Maxwellian-view image directly onto the retina, giving a focus-free picture; the product line has been used both as a consumer viewer and as a low-vision aid for people whose eyes cannot focus normally. Published descriptions report a horizontal field of view of about 25 degrees, around WSVGA resolution, a 60 Hz update rate, and laser power low enough to satisfy the Class 1 safety criterion.

See also

• Near-eye display
• Head-mounted display
• Augmented reality
• Microdisplay
• Maxwellian view
• Field of view
• Waveguide display