Advancing free-space optical communications with adaptive optics
Eliminating optical distortions to improve free-space-optic communications has astronomical origins.
J. ELON GRAVES and STEVE DRENKER, AOptix Technologies
Network operators are seeking ways to cost-effectively break the metro bottleneck and deliver high-capacity voice and data to bandwidth-hungry customers. The distribution of high-capacity fiber directly to customer buildings requiring greater bandwidth than the existing copper network can provide is cost-prohibitive and will take decades to accomplish. Nationally, estimates put fiber terminations in only 5-10% of medium-to-large office buildings, but three-quarters of these buildings are located within one mile of fiber.
Closing this "bandwidth gap" is the goal of both radio- and optical-frequency wireless communications technologies. Fixed broadband wireless systems are limited in bandwidth and require heavy capital outlays to secure spectrum licenses and deploy networks in advance of paying customers. Wireless optical systems, also known as free-space optics (FSO), address several of these issues by transmitting at unlicensed optical frequencies and using point-to-point architectures that are truly "pay-as-you-grow," avoiding heavy up-front capital outlays.
But FSO transmissions are susceptible to atmospheric-induced attenuation. System designers have tried to overcome this problem in a variety of ways-with only limited success. A new solution, adaptive optics (AO), can restore lightwaves deformed by the atmosphere to near-original condition, tackling the root-causes of scintillation, wander, and beam spreading. Adaptive optics promises to revolutionize FSO communications by meeting new carrier requirements: true speed- and protocol-independent links that operate over longer distances, at higher bandwidths, through any office window, and with greater reliability-in all weather conditions.
The idea of using optical signals transmitted through air to exchange information between two locations is not new. Alexander Graham Bell transmitted the first wireless telephone message in June 1880 over his newly invented "photophone," projecting voice through an instrument toward an acoustically coupled mirror. Voice vibrations set up similar vibrations in the mirror, which reflected and projected sunlight toward a receiver where the vibrations were transformed back into sound. Commercial success eluded the photophone because it was susceptible to outside noise and transmission over copper wire was more reliable.
It wasn't until 80 years later, when the laser was developed and demonstrated that the use of modulated light for through-air communications became practical. Early development began in the 1970s by the United States for secure aircraft-to-ground and satellite-to-submarine communications. Commercial FSO systems became available in the 1990s, primarily to interconnect buildings and campus LANs, and for such specialty applications as field-site TV camera communications. These markets are generally characterized by short ranges, modest data rates, and outdoor installations. Systems built for this market are made from commercial, off-the-shelf optical components such as lasers, receivers, and lenses.
Recently, telecommunications carriers began to look for FSO systems to meet the increasingly demanding requirements found in new markets, including last-mile access, metro ring extension, and wireless backhaul. Customer requirements in these markets include:
- Longer range and greater reliability in all weather conditions.
- True physical layer data-rate and protocol independence.
- Carrier-grade products with integrated network-element management.
- Outdoor and indoor operation through virtually any office-window glass.
- Single-channel data rates up to 10 Gbits/sec and beyond with WDM.
Why adaptive optics?
To meet the market requirement for bit-rate and protocol independence, FSO systems would ideally operate at the physical layer, and the optical transmission system would "recognize" the free-space optical system as a piece of fiber. Instead of using protocol- and data-rate-dependent laser transmitters and receivers, it is desirable to take light directly from a transmission fiber, amplify it, transmit it through the air, capture it at the receiver, and focus it onto the 9-micron core of the receiving singlemode fiber. Because of low energy levels in the beam, virtually all of the optical energy emitted from the transmitter needs to be captured by the receiver to deliver an acceptable signal-to-noise ratio. Therefore, the beam needs to be narrow and highly collimated.
Collimated transmission beams have not been used in previous-generation FSO systems for several reasons. First, atmospheric distortions cause the beam to "wander" out of the receiver aperture, leading to loss of signal. Beam wander is caused by the non-uniform index of refraction in turbulent air between the transmitter and receiver, a result of fluctuating air temperature, pressure, and wind (see Figure 1).
Pockets of air with varying densities exist in a range of sizes-from a few millimeters to meters in scale-and they constantly grow, shrink, and move around at rates up to 100 Hz. Because the speed of light through air depends on the index of refraction, these ever-changing pockets of air act as a collection of changing lenses that cause the beam to wander at the receiver. Sources of heat that contribute to wander are common in urban and suburban areas and include streets, parking lots, rooftops, and building exhaust vents.
Second, the constantly changing index of refraction causes some parts of the beam to slow more than others, distorting uniform wavefronts that exited the transmitter. These small, random phase changes cause constructive and destructive interference (see Figure 2).
Third, changes in index of refraction cause beams to spread out in transit, reducing the energy on the central axis. Finally, beams wander because of the relative motion or sway of buildings and towers exposed to wind loads and non-uniform thermal expansion from solar heating.
This collection of challenges has been historically addressed by using:
- Highly divergent beams that spread their energy over a wide diameter at the receiver. That ensures the receiver aperture is always illuminated, regardless of beam wander and structure sway (see Figure 3) but at the cost of greatly reduced energy incident upon the receiver.
- Higher-power transmitter lasers to overcome losses caused by beam divergence and variations in received light intensity caused by scintillation.
- Multiple transmitter beams, which pass through different air pockets and are subjected to different "atmospheric lenses," thus increasing the probability that at least one beam will not be impaired, while increasing average received power. The benefits of multiple beams are greatest near the transmitter, where beams are most physically separated. As the beams begin to merge some distance from the transmitter, they once again pass through the same atmospheric pockets, experiencing the same wander and scintillation effects.
- Larger receive apertures, so more rays are captured, reducing the variation in scintillation.
- Decreased maximum link distances to simply avoid some atmosphere.
Even when combining several of these solutions, it is difficult to capture enough energy to produce a reliable all-optical solution, because the fundamental problem is not addressed: the wavefront of the light is distorted by the atmosphere. That opens the door to the use of another more effective countermeasure: adaptive optics.
Many frequent fliers know that noise-control headphones are available that greatly reduce aircraft noise during flight to make a trip much more comfortable. These headphones work by creating "anti-noise," a reversed-phase signal that cancels out cabin noise. In a similar fashion, AO is a way to apply "anti-noise" signals to light beams to cancel out noise (aberrations) that impair light as it passes through turbulent atmosphere and office-window glass.
AO was first developed in the late 1950s to correct atmospheric blurring by tilting a telescope's secondary mirror several times a second. That reduced the wander of the image and improved image sharpness twofold. However, a tenfold improvement was needed to completely eliminate atmospheric blurring in astronomical observations.
The military invested heavily in AO systems in the '70s and '80s to deliver concentrated laser energy onto enemy targets. When advanced military technology was later declassified, astronomers quickly adopted it to totally eliminate atmospheric blurring of star and galaxy images. Custom AO systems used in the world's largest telescopes are hand-made, cost hundreds of thousands of dollars, and take many years to produce. The results, however, are astonishing. Astronomical images taken from earth observatories equipped with AO rival the quality of those taken by the Hubble space telescope from outside the earth's atmosphere. This use of AO is the most revolutionary development in astronomy since Galileo invented the telescope in 1609.
AO systems unite several advanced technologies to remove noise from optical signals (see Figure 4). Light first passes through an objective lens and is then reflected off a special adaptive optical element called a deformable mirror that applies the anti-noise to the beam, canceling aberrations on the incoming beam. Part of the corrected wavefront is split off to a wavefront sensor that measures residual distortion or error in the wavefront. A correction signal is calculated in a CPU and sent to the deformable mirror. The light not deflected by the beam splitter continues to a receiver, shown in Figure 4 as a bidirectional fiber-optic data port.
Time delay, or latency, between the measurement of residual wavefront error and the movement of the deformable mirror must be kept to a minimum, otherwise the system will correct for past, not current, atmospheric conditions. That is challenging, with rapidly changing random processes such as atmospheric turbulence, where the future state cannot be inferred from current conditions. In such situations, the control system must operate at least 10 times faster than the process being controlled. Since the atmosphere is changing on the order of 100 Hz, the control system must make corrections in the deformable mirror at 1,000 Hz or faster.
Previous military and astronomical AO systems have been unidirectional, either correcting deformed light from stars and galaxies through telescopes or pre-deforming outgoing light in the case of laser weapons. FSO systems are now available for the first time that point two AO units at each other in a bidirectional communications system. The deformable mirror therefore not only corrects the incoming-light wavefront, but also simultaneously pre-deforms the outgoing-light wavefront to pre-correct for known aberrations in the light path between the two units. Pre-shaping the beam from the transmitter to correct for known atmospheric conditions eliminates wander and scintillation so that the beam can be locked onto its target.
Simulations were conducted to compare the performance of all-optical systems operating over a 2-km link distance with and without adaptive optics. Without AO, scintillation causes fast, deep signal fades (see Figure 5a). By applying AO, the average scintillation loss and the standard deviation of loss are greatly reduced (see Figure 5b). By correcting the fundamental wavefront distortions that cause beam wander, scintillation, and spreading, this system is able to transmit highly collimated beams, capture all the transmitted light at the receiver, and couple it into a 9-micron singlemode-fiber core.
Addressing root problems
An FSO system using AO challenges current design trends that target improved system performance with more and higher-power lasers combined with creative packaging of standard optical components. AO has many inherent advantages in addressing the root causes of FSO transmission impairments.
It is now possible to transmit bidirectional, highly collimated beams that overcome scintillation by correcting deformed wavefronts thousands of times a second. As a result, it is also possible to transmit directly from a customer's singlemode fiber through window glass and air and capture the light back into the receiving fiber, all with higher link margins that deliver longer range, more reliable links, and higher bandwidth.
J. Elon Graves is founder and vice president of optics and Steve Drenker is director of product management at AOptix Technologies Inc. (Campbell, CA). They can be reached via the company's Website, http://www.aoptix.com.