Free-space communications struggles to find its niche

The vast majority of today's optical communications runs through fibers, while the vast majority of free-space wireless communications is transmitted at radio wavelengths. Producers of free-space optical communications have had an uphill battle to find markets that are not better served by either fiber or radio. So far, only a relatively small portion of the market for interbuilding short-distance communication has been penetrated by wireless optical in areas where, for some reason, laying fiber is too expensive.

To expand substantially, free-space optical will have to prove itself in larger markets—either the "final mile" for urban users, indoor mobile communications, or for space communications. But in all these areas, free-space faces significant technical challenges that have only begun to be overcome, as well as tough competition from other technologies.

The interbuilding niche

Free-space optical communications has been limited to links between buildings at headquarters, campuses, or similar facilities (see Laser Focus World, November 2001, p. 101). The key advantage of such interbuilding optical links is that the costs of cabling for fiberoptics are avoided, as is the inconvenience of getting a license for microwave communications. In addition, free-space optical links have the same huge bandwidth as fibers, far beyond the capabilities of microwave links. Typical commercial equipment carries 1 Gbit/s, and experimental devices have transmitted as much as 160 Gbit/s over a 200-m link.¹

However, optical free-space links are far more affected by the weather than are microwave links. Fog in particular can severely interfere with free-space communications; for high reliability a microwave backup is needed, considerably reducing the advantages of optical links.

Into the upper atmosphere

One way to avoid the weather problem is to get above it and look to applications in the upper atmosphere and space. The U.S. Air Force Research Laboratory (Wright Patterson AFB, OH) is currently testing an air-to-air link designed to transmit 1 Gbit/s over ranges up to 500 km at altitudes of 40,000 ft.² The need for wide-band communications arises from the use of sensors like synthetic-aperture radar, which generates huge amounts of data. Storing the data on an aircraft requires bulky electronics, so transmitting them to another aircraft for immediate download to an Earth-based station is desirable.

The air-to-air link's laser terminal is mounted in a turret on the underside of a T39-A test aircraft (see Fig. 1). The optical subsystem in the turret consists of a communication and coarse-tracking telescope with array sensors, two beacon lasers, a fine-tracking telescope and sensor, four communications lasers, and a fast-steering mirror, all attached to a composite optical bench. Initial alignment is established by transmitting the Global Positioning system positions of the two aircraft to each other by radio. This allows the coarse-tracking telescope to lock onto the broad coarse-tracking beacon. The fine-tracking telescope then locates and locks on the fine-tracking beacon, keeping the very fine communications laser beam within 12 μrad of the communications receiver. The four lasers combine to form two channels, each with a 600-Mbit/s data rate and a 200-mW output. Flight testing of the system will be carried out throughout this year.

FIGURE 1. An Air Force test air-to-air laser communications terminal is mounted in a turret under a test aircraft. It includes laser beacons for both coarse and fine tracking, four laser communications diodes, and tracking and communication sensors, all fed through a 20-cm telescope.

Easing into space

Although free-space optical communication in space avoids the problem of weather, the pointing and tracking problems get worse. One of the key problems is acquisition of one satellite's signal by another satellite. For initial acquisition, a two-dimensional scanning pattern must be used, starting with the assumed location of the other satellite. An ideal scanning pattern would concentrate more time on the most likely position of the satellite, while covering the whole possible area rapidly. Studies at Ben Gurion University (Beersheba, Israel) have shown that a rose or lissajous pattern tends to be the most efficient in finding a laser beam (see figure at top of page).³ Savings in time can be quite substantial. For a Gaussian distribution of beam locations around a central position, rose and lissajous patterns take less than half as long to find the beam as does a simple raster scan.

The first telecommunications satellite to use free-space lasers, Artemis, is now in orbit after barely escaping disaster. The European Space Agency satellite was launched by an Ariane 5 rocket in July 2001 but failed to reach a proper orbit as a result of a rocket malfunction. Fortunately, the satellite's new ion rocket, originally designed to keep it in place in geosynchronous orbit, came to the rescue and was ordered to boost the satellite up to geosynchronous.

Artemis is now linked to another satellite, the Earth Observation Satellite Spot 4, over a 60-mW laser operating in the 800- to 850-nm band with a 50-Mbit/s capacity. The beam width of only a few microradians makes accurate pointing and tracking essential.

Inner space

Another potential area of market expansion is inside-building networks, in which again the weather does not matter. Here, optical free-space networks face tough competition from the Bluetooth network developed by IBM (Armonk, NY), which automatically links electronic equipment via a very-short-range radio signal. The communications systems of many users are now Bluetooth compatible, making it more difficult for alternative technologies to compete.

However, indoor wireless optical communications based on infrared emitters and sensors has advantages. It has a much larger bandwidth (important for downloading video data), it is immune to electromagnetic interference, and it offers better security because it cannot pass through walls. So far, these advantages have not been enough to overcome the big Bluetooth lead.

To further open the bandwidth gap between optical and radio-frequency (RF) indoor systems, engineers are developing new modulation schemes that more efficiently use the available bandwidth. For example, pulse-interval modulation encodes information in the length of interval between a short-duration pulse at the beginning of each frame and subsequent information-carrying pulses.⁴ These new schemes can substantially reduce the power required for a given signal-to-noise ratio and information-carrying capacity.

One of the key disadvantages of optical free-space links compared with RF—even in an indoor environment—is that optical links are line-of-sight and bidirectional while RF links offer one-to-many or many-to-many and are omnidirectional. To avoid this problem, engineers have been developing quasi-diffuse links. Instead of transmitting a single beam, transmitters emit multiple narrow beams that illuminate small areas, or diffusing spots, on a diffuse reflecting surface such as a wall or ceiling. Receivers also have many elements to receive from more than one diffusing spot, so that communication is uninterrupted even if some of the transmitter beams are blocked by, for example, someone walking in front of them.⁵

Multiple diffuse links create multipath distortion, because there are several paths of different lengths for signals to get from transmitter to receiver. One way to minimize this problem, which can degrade signals, is to ensure that the diffusion spot pattern is matched to the receiver field of view in such a way that only one spot is within each field of view (see Fig. 2).

FIGURE 2. Creating many diffuse spots (small dots) on a wall or ceiling can prevent free-space communications in a building from being interrupted. To avoid the problem of multipath distortion from many spots, the field of view of individual receiver branches (large circles) can be adjusted for that; at most one diffusion spot lies within each field of view (right).

In brightly lit offices, background optical noise is also a problem. Receivers for optical communications need narrow spectral filters and optical concentration to achieve good signal-to-noise ratios. One approach to this problem is to use a holographic optical element for both filtering and light concentration. A holographic mirror will act as a spherical mirror to concentrate the light and will only reflect to the detector a narrow spectral range. Such concentrators can improve signal-to-noise ratios by up to 18.5 dB, allowing the use of bandwidths up to 2 GHz.

The commercial future of free-space optical communications remains uncertain. On the one hand, the push for ever-greater bandwidth could eventually outrun any RF capacity, forcing a shift to free-space optical. On the other, RF techniques have a long head start in inside-building applications, and the weather problem remains a severe one for optical interbuilding links. Perhaps the best overall prospects are in space, where progress is being made in improving acquisition and tracking. Once these are perfected, the bandwidth advantages of optical free-space communications should open up a substantial market niche.

REFERENCES
1. G. Nykolak et al, SPIE 4214, 11 (July 2001).

2. J. Petrovich et al, SPIE 4214, 14 (July 2001).

3. N. S. K. Sheinfield and S. Arnon, SPIE 4365, 195 (September 2001).

4. N. M. Aldibbiat et al, SPIE 4214, 144 (July 2001).

5. S. Jivkova and M. Kavehrad, SPIE 4214, 162 (July 2001).

EDITOR'S NOTE: Optical communications is the subject of this month's "Focus On . . ." series. In addition to this article on free-space communications, several other features address different aspects of communications: "Optical Networking" discusses optical packet switching on page 131, a look at erbium-doped waveguide amplifiers is on page 101, and optical performance monitoring is covered on page 149.