JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 23, NO. 6, JUNE 2005

2023

Backscattering-Induced Crosstalk in WDM Optical Wireless Communication Debbie Kedar, Student Member, IEEE, and Shlomi Arnon, Senior Member, IEEE

Abstract—The crosstalk effect of aerosol backscatter on the performance of a wavelength-division-multiplexed (WDM) optical wireless communication (OWC) system is investigated, analyzed, and quantified. An OWC link could be a segment within a metropolitan area network (MAN) or a ground-station-to-space link of a satellite communication system. In these cases, a WDM transmitter and receiver are housed in one transceiver unit with parallel, or near-parallel, optic axes. The crosstalk at the receiver is caused by light from the transmitted signal of the same transceiver, which has been backscattered by molecules and aerosols in the atmosphere. This is exacerbated in the presence of fog and haze, in which case both the desired signal from another transceiver is attenuated by scattering and the backscatter-induced crosstalk increases. A bit-error-rate (BER) model is derived that takes into consideration the dominant noise sources, including backscatter-induced crosstalk and signal mixing with amplified stimulated emission (ASE) from an optical preamplifier at the receiver. The numerical calculations in this paper indicate that, in moderate fog, the BER may increase by an order of magnitude or more due to backscatter, depending upon the atmospheric extinction coefficient. Index Terms—Amplified spontaneous emission (ASE), backscattering, crosstalk, Mie scattering, optical preamplifier, optical wireless communication (OWC), Rayleigh scattering.

I. INTRODUCTION

O

PTICAL communication—in space and in both outdoor and indoor terrestrial applications—is rapidly becoming a major feature of modern life. While optical fiber communication has matured over the past decade, optical wireless communication (OWC) is still considered an emerging technology and is being widely researched, particularly with regard to signal degradation caused by propagation through the atmosphere [1]–[4]. In an urban setting, the growth in demand for high data rate transmissions, especially Internet applications and company intranet systems, has led to the expansion of metropolitan area networks (MANs) [5], [6]. Optic fiber communication can provide the extremely high bandwidth required, but few office buildings and business premises are connected to the optical fiber backbone. Wavelength-division-multiplexed (WDM) OWC can seamlessly bridge the gap, and hence relieve the severe bottleneck otherwise encountered over “the last mile” to the client premises from the fiber backbone. The major source

Manuscript received July 12, 2004; revised February 3, 2005. This work was supported by the DIP (Deutsche-Israelische Project) Fund within the BLISL project. The authors are with the Satellite and Wireless Communication Laboratory, Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev, IL-84105 Beer-Sheva, Israel (e-mail: [email protected]). Digital Object Identifier 10.1109/JLT.2005.849875

Fig. 1.

Schematic illustration of the communication scenario.

of atmospheric degradation in OWC, both in an urban setting and in satellite communication, is multiple scattering by particles in the propagation channel. Link failure due to multiple scattering is frequently caused by adverse weather conditions, particularly fog and haze [7]. In these circumstances, not only is the desired signal attenuated, but backscattering of transmitted light is increased. Stray scattered light from one transceiver may erroneously reach the receiver of another transceiver, where it becomes crosstalk interference and obscures the desired signal, reducing the signal-to-noise ratio (SNR). When the light is backscattered from a transmitted beam to the receiver of the same transceiver (see Fig. 1), the overlap between the propagating beam and the receiver field of view (FOV) may be considerable (see Fig. 2), and a great deal of crosstalk may be encountered. Crosstalk interference is one of a number of noise sources at the receiver. Optical preamplifiers at the receiver are a common means employed to increase effective receiver sensitivity and to reduce the deleterious effect of receiver electronic noise. Unfortunately, optical preamplifiers also contribute to the overall system noise and amplified spontaneous emission (ASE) beats with the other power sources reaching the receiver, such as backscattered light from the transmitter. The phenomenon of crosstalk has been extensively researched, especially in the context of optical fiber communication, where it is a major capacity limiter in WDM networks [8]–[14]. Crosstalk between network users has also been investigated in the context of free-space optics [3], [15]. However, to the best of the authors’ knowledge, no work has been published dealing with crosstalk due to backscattering between the transmitter and receiver of a single transceiver. In this paper, we analyze the influence of backscattering-induced crosstalk noise in a -band (1530–1565 nm) WDM transceiver and quantify the phenomenon in an urban setting.

0733-8724/$20.00 © 2005 IEEE

2024

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 23, NO. 6, JUNE 2005

and haze, scattering is the predominant mechanism that influences light propagation [16]. Scattering refers to energy conservative deflection of propagating light from its original direction. When the scattering object is a gas molecule or a very small particle, the scattering mechanism can be described by Rayleigh theory. The propagating light is deflected multidirectionally with intensity inversely proportional to the fourth power of the wavelength of radiation. The amount of light scattered in any given direction is described by the phase function, or probability distribution function of angle of scatter, which has been tabulated in the literature [17]. When the scattering object is an aerosol or water droplet of radius similar to the radiation wavelength, Mie scattering occurs. The Mie phase function is more directionally complex than the Rayleigh phase function but can be derived analytically once the particle size distribution has been defined. However, standard tables of backscattering coefficients, for a variety of radiation wavelengths, geographical locations, altitudes above ground and sea level, and ambient weather conditions are used in this work [17], [18]. Fig. 2. Transceiver showing (a) transmitted beam, with a small divergence angle, described by ' , positioned above the receiver FOV cone, described by ' , and (b) the deviation-from-parallel of the transmitter beam and FOV, described by  and  .

WDM optical wireless transmission facilitates seamless connection with WDM fiber systems. Today, WDM is the preferred multiplexing method in optical communication and the technology for its operation is continuously maturing. In WDM systems, multiple users can be supported simultaneously; networking further increases traffic by expanding interconnectivity between nodes, particularly when individual links are inoperable. Clearly, in a dedicated point-to-point link where the capacity requirements do not necessitate using all available wavelengths on each side of the link, different wavelengths may be assigned for transmitting and receiving. In this case, backscatter-induced crosstalk would not be encountered. The remainder of this paper is organized as follows. Section II briefly reviews atmospheric scattering, and Section III describes backscattering-induced crosstalk and other noise sources using a mathematical model. In Section IV, a numerical example is presented, and the last section discusses the results and outlines conclusions. II. ATMOSPHERIC SCATTERING A. Rayleigh and Mie Scattering Our planet Earth is swathed in an envelope of atmospheric gases and particles. The interaction of propagating light with this matter can be described by mechanisms of absorption, scattering, and turbulence. Light is absorbed in a highly wavelengthsensitive manner, resulting in attenuation. Turbulence is an important time-varying phenomenon encountered when the optic path passes through pockets of varying temperatures and results in scintillation. However, turbulence effects are not usually considered significant when short propagation distances (of less than one kilometer) are considered and when the logarithm variance of the scintillation is small [2], [3]. In the presence of fog

III. MATHEMATICAL MODEL We now develop a mathematical model to describe the optical communication link. We formulate the backscatter power on the basis of the bistatic lidar equation, using backscattering coefficients from standard tables, and define the noise sources at the receiver. An expression for the bit-error rate (BER) is derived. A. Link Budget In this paper, we consider the transceiver of an OWC link using intensity modulation and direct detection (IM/DD). To find the relation between the received and transmitted power in a communication link, we define the power link budget as (1) where is the received optical power and is the transand are mitted power of the desired communication link; is the atmothe transmitter and receiver optical efficiencies; spheric transmissivity defined in terms of the optical density of the propagation medium (2) and is the ratio between the receiver aperture area and the area of the divergent beam spot at the receiver. The essential geometry of the transceiver is shown in Fig. 2. The beam divergence and and the receiver FOV are cones described by the angles , respectively. The initial beam diameter and receiver aperand , respecture diameter (not shown in the figure) are tively, from both communicating transceivers. We approximate the energy distribution within the beam as near uniform. If the distance between the two communicating transceivers is (the receiver in one transceiver receiving an intended signal from the transmitter in the other), we define (3)

KEDAR AND ARNON: BACKSCATTERING-INDUCED CROSSTALK IN WDM OWC

B. Backscattered Optic Power We now examine the geometry of the transmitter beam and the receiver FOV in more detail (Fig. 2). The optic axis of the transmitter is located at a distance from the optic axis of the receiver in the transceiver housing (in Fig. 2, the transmitter is vertically above the receiver). In the case when the axes of the beam and of the receiver FOV cones are parallel, and [as in Fig. 2(a)], once the beam enters the FOV cone at a distance from the housing, the two cones continually intersect. Moreover, from distance until infinity, the intersection is total. From simple geometric considerations, the values of and can be given by (4) (5) , the values of and In the case when are the same as the values given in (4) and (5), exand , and likewise and , are incept that terchanged. In the case when , then and . In a real transceiver, the axes of the beam and the FOV are not necessarily perfectly parallel to ensure that the transmitter will be aligned with its intended receiver, and vice versa. In Fig. 2(b), the deviation-from-parallel angles are described by and , respectively. If , the two cones intersect. The values of and are the same as those is replaced by given in (4) and (5), except that the angle , where . The transmitter beam will start to exit the FOV cone at a distance from the transceiver given by

If and totally exit the FOV cone at

(6) , then the beam will

(7) It is clear from Fig. 2 that, for small deviation-from-parallel angles, the intersection volume of the two cones will be similar to the case where the beam and FOV cones are parallel . The power erroneously received by the receiver from the transmitter of the same transceiver due to backscattering is given by the bistatic lidar equation [19]

(8) where is the power in the transmitter beam of the same transceiver, is the distance along the beam from the transmitter, is the time measured from the signal transmission initiation, is the velocity of light, and describes

2025

the transmitted power delayed in time by twice traversing a disfrom the transceiver to the backscattering particles tance of is further defined by at . (9) where duration , and

represents the value information bit of represent a pulse shape given by (10) .

and are the backscattering cross sections for molecules and for aerosols, respectively, per unit atmospheric is the receiver’s length at distance from the transmitter; spectral transmission factor, including the influence of any is the geometric form factor spectrally selecting elements; and relates to the spatial intersection of the transmitter beam and is the extinction coefficient; and the receiver FOV cone; is the receiver aperture area. The sum of the two backscattering cross sections expresses the combined backscattering effect of describes the atmosphere at . The term as the round-trip attenuation encountered by the signal it propagates to and back, which comprises the backscatter is the acceptance solid angle at the signal attenuation. receiver for backscattered light from a particle at distance , . The molecular backscattering cross section where is further defined as [20] (11) where is the differential Rayleigh backscattering cross section per “average” gas molecule in units of is the number of gas molecules per unit m sr , and volume, which is assumed to be m in the atmospheric boundary layer. For the mixture of the atmospheric gases that occurs below about 100 km (12) where is in micrometers. is defined as the degree of The geometric form factor overlap between the illuminated area and the scattering area as seen by the receiver aperture. In the work reported in this paper, we assume that the FOV angle, while small, is nevertheless considerably larger than the laser divergence, since both a small divergence angle and a large FOV minimize pointing errors in transceiver alignment. Hence, the appropriate expression is (13), shown at the bottom of the next page [20], where for and are the total beam diameter and FOV diameter at and are given by distance from the transceiver, and (14) (15)

2026

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 23, NO. 6, JUNE 2005

for the case when the beam and FOV cones axes are parallel. When angular deviation-from-parallel is added, the expressions and are the same as those given by (14) and (15), for is replaced by . except that If the optical link is approximately horizontal, at altitude , then the backscattering cross sections and the extinction coefficient can be assumed to be constant and are no longer a function of . Consequently, the sum of the two backscattering cross secand tions can be written outside the integral in (8), as can , which are also not dependent on . Assuming that binary “0” and binary “1” are equiprobable, we can now express the time-averaged backscattered power as a function of the time-averaged transmitted power as

channel as having a memory of one symbol. (Note that the backscatter reaching the receiver from previous symbols, and consequently from more distant parts of the intersection cone, would be significantly attenuated by propagating through the additional length of propagation path.) We therefore describe four permutations of signal and backscatter: 1) signal and backscatter ; 2) the reverse; 3) both “1”; and 4) both “0”. The noises accompanying each of these four permutations , , , and , respectively. Substitution of are termed realistic values for all noise terms (assuming low background illumination) shows that optical amplifier noise dominates all other sources. Hence, the noise current variances are given by [21]

(16) (17) C. Signal and Noise In this section, we outline the major noise sources that obscure the signal at the receiver. All received optic power is amplified by an optical preamplifier with gain and noise factor and then converted to an electronic current by a photodiode , where is the with responsivity given by quantum efficiency of the photodiode, is the electron charge, is Planck’s constant, and is the optical frequency of the received power. Within the narrow range of the band, we as, which sume the responsivity is constant. Likewise, and also vary with optical frequency, will be assumed constant in between the this paper. An optical bandpass filter of width preamplifier and the photodetector restricts the optical band, width of the detected power. The desired signal power is while all other optic power received is interference and obscures the signal with additive noise components. At the receiver, the backscattering-induced crosstalk is one of a number of interferences that degrade the signal that is received from another transceiver (e.g., signal shot noise, dark current noise, background noise, thermal noise, and ASE mixing noise). We consider all these noise sources, including the backscattering-induced crosstalk, as independent and Gaussian and assume that the modulation technique for both transmitters is ON–OFF keying (OOK) (as stated in the section on the link budget). Consequently, the transmitted signal is either a binary “0” or a binary “1” both in the intended and in the backscattered signal. The noise accompanying each of these two possibilities is not the same in the presence or absence of backscatter (when the undesired locally transmitted signal is “1” or “0”, respectively). Henceforth, for simplicity, we model the backscatter

, , and are the noise terms where due to beating of the amplified spontaneous emission (ASE) from the optical amplifier with itself, the signal, and the is the noise backscatter-induced crosstalk, respectively. due to the mixing of the signal with the backscatter. The beat noise terms caused by the mixing of two signals can be defined as [21] (18) (19) (20)

(21) (22) where is the electronic bandwidth. The last term (22) requires some explanation. In contrast to the ASE mixing noise terms, which have a spectrum as broad as the optical bandwidth of the amplifier, the backscatter-signal crosstalk originates from the mixing of two sources at the same frequency. The two signals are not coherent; therefore, the resulting interference does not take the form of simple amplitude fading as would be caused by the addition of two coherent signals of randomly varying amplitude. To obtain the signalbackscatter crosstalk, it is necessary to compute the time-averaged product of the two current signals, which differ by a ran, which can vary domly and rapidly changing phase shift

(13)

KEDAR AND ARNON: BACKSCATTERING-INDUCED CROSSTALK IN WDM OWC

between- and . (Since more than photons arrive simultaneously from many different scattering locations within the overlap volume per nanowatt of backscattered power, the phases of the interfering backscattered light would be totally random in time and space relative to the desired signal.) In clear weather, the background radiation and, in consequence, the ASE background beat noise may be significant, but then the backscattered light will be relatively low and the ASE backscatter beat noise can be ignored. Conversely, in the presence of haze or fog, the backscattered light will be significant, and the background radiation negligible in comparison. Hence, the ASE background beat noise can be ignored in these cases.

2027

that backscatter from previous pulses is highly attenuated. We now expand (24) to render signal

backscatter

signal

backscatter

signal

backscatter

signal

backscatter

D. Bit-Error Rate The total received optic power is converted to an electronic signal by the photodiode. With the addition of receiver noise . It is the task of sources, the electronic signal is termed , the receiver decision circuitry to determine, on receipt of whether the desired signal transmitted from a distant transceiver with a threshold was “1” or “0”. This is done by comparing , the received signal is interpreted as a current . If , it is interpreted as a transtransmitted “0”, while if mitted “1”. We term the a priori probabilities of binary “1” and “0” being transmitted in the first place as P(“1”) and P(“0”), respectively. Furthermore, we introduce the conditional probability and , which are functions density functions would be received if the describing the probability that transmitted signal was “1” or “0”, respectively. We then can derive a general expression for the BER, modeling all accompanying noise as independent and Gaussian by calculating the probabilities of the two possible errors—that a “1” is sent and is interpreted as a “0”, and the reverse. Hence, the BER is formulated as BER

(23)

We define the average received current from the signal and from the backscattered interference as and , respectively, when a “1” has been received, and zero when a “0” is received (zero extinction ratio). Averaging the cumulative signal over their phase difference, we can expand (23). We recall that transmissions of binary “0” and binary “1” were considered to be and can now write equiprobable and as signal signal signal signal

backscatter backscatter backscatter backscatter

(24)

We model the extreme case where the backscatter and the signal are synchronized, and only relate to backscatter ensuing from a single pulse from the local transmitter on the assumption

(25) Substituting (25) in (24), we get

(26) Defining the complementary error function, as , we can now summarize the BER as (27), shown at the bottom of the next page. IV. NUMERICAL CALCULATION AND RESULTS In order to evaluate and quantify the influence of the backscattering-induced crosstalk, we present a numerical example (see Table I). The backscattering coefficients for haze and fog are derived from tabulated data in the literature [17], [20]. For our m m and purposes, we have used 7 m sr as typical values for moderate fog. The results of our numerical calculations are summarized in Fig. 3. The BER is plotted as a function of optical density (be. (The optween 7 and 10) for two different values of FOV, .) To demonstrate the deleterious tical density is given by effect of the backscatter, a third trace is plotted showing the resultant BER for the case when the backscatter optic power is not taken into account. We observe that between an optical density of 7 and 8, the backscatter power significantly degrades the communication system performance. If the backscatter interference is not taken into consideration, then the anticipated BER for the

2028

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 23, NO. 6, JUNE 2005

TABLE I VALUES FOR VARIABLES USED IN THE NUMERICAL EXAMPLE

system is underestimated by a factor of 3 for an expected BER when the FOV is 1.5 mrad. When the anticipated BER of , the corresponding underestimation is by a factor of is 4. As the optical density increases the desired signal is increasingly attenuated, which results in high BER values regardless backscatter. In an OWC system, it is often preferred to operate with an increased FOV in the presence of fog and haze in order to increase the received power, by including more multiply scattered light [7]; however, an increase in the FOV could result in increased channel memory. Note that in our example, the BER and by a factor increased by a factor of 40 for a BER of , when the FOV is increased from 1.5 of 20 for a BER of to 3.0 mrad. The increase in BER when the FOV is larger is due to the increased proximity to the transceiver of the initial point of intersection of the beam and FOV cones. As a result, much of the backscattered light reaching the receiver will have undergone very little attenuation. (The increase in received power due to the enlarged FOV is negligible since the receiver–transmitter pair is assumed in (1) to be perfectly aligned [7].) The phenomenon mentioned in the last paragraph is also evident in Fig. 4, which shows the increase in BER with deviation-from-parallel angle [see Fig. 2(b)], for the case when 1.5 mrad and transmitted power is 5 mW. A deviationfrom-parallel angle of 1.0 mrad increases the BER by three orders of magnitude by comparison to the parallel axis configuration for an optical density of 7. For optical densities of 7.5 and

Fig. 3. Graph of BER versus optical density  for two values of receiver FOV, showing an increase in BER for the larger FOV. BER without inclusion of backscatter (but with atmospheric attenuation) is shown for comparison (solid line).

Fig. 4. Graph of BER versus deviation-from-parallel angle  , showing an increase in BER with  for three different optical densities. The receiver FOV angle is 1.5 mrad, and the transmitted power is 5 mW.

8, the BER increases by factors of 27.4 and 3.3, respectively, when the deviation-from-parallel angle is 1.0 mrad. Finally, we analyze the results of the numerical example in terms of power penalty (see Fig. 5). We stipulate a minimum

BER

(27)

KEDAR AND ARNON: BACKSCATTERING-INDUCED CROSSTALK IN WDM OWC

2029

In summary, neglecting the influence of backscatter-induced interference in analysis of OWC systems under conditions of fog may result in misleading predictions of performance. ACKNOWLEDGMENT The authors are grateful to the DIP Fund (DeutscheIsraelische Project Fund) for support within the BLISL project. REFERENCES

Fig. 5. Graph of power penalty versus optical density for BER of 10 . FOV is 1.5 mrad. The power penalty without inclusion of backscatter (but with atmospheric attenuation) is shown for comparison (solid line).

BER of and compute the power penalty for the same range of optical densities studied previously, with an FOV of 1.5 mrad. This corresponds to a transmitted power of 3 mW under conditions of an optical density of 7. The power penalty due to atmospheric attenuation alone, without inclusion of backscatter interference, is shown in Fig. 5 by the solid line for comparison with the results including backscatter. In order to maintain the stipulated BER, the power penalty rises from zero to 17 dB as the optical density increases from 7 to 10. If backscatter degradation is not considered, the power penalty rises nearly linearly from zero to 13.4 dB over the same optical density range. This results in an increasing discrepancy between power penalties with and without backscatter as the optical density rises, reaching a discrepancy of over 3.6 dB at an optical density of 10.

V. CONCLUSION In this paper, it was shown that crosstalk noise caused by the backscattering of transmitted light from a transceiver onto the receiver of the same transceiver can reach levels that significantly degrade the communication performance. The crosstalk noise increases with the FOV, since the overlap between the beam and the FOV begins nearer the transceiver, and hence the backscattered power that arrives at the receiver is minimally attenuated. Similarly, deviation-from-parallel of the transmitter and receiver axes in a transceiver increases crosstalk noise when the laser beam enters the FOV nearer the transceiver, resulting in more backscattered light reaching the receiver. Divergent deviation-from-parallel angles would be expected to reduce the backscattering-induced crosstalk noise. Furthermore, it is concluded that if the backscatter is not considered when calculating the power penalty in optical wireless communication (OWC) transmissions through mild fog, the error in the calculation of power penalty will increase as the optical density rises.

[1] Y. Alqudah and M. Kavehrad, “MIMO characterization of indoor wireless optical link using a diffuse-transmission configuration,” IEEE Trans. Commun., vol. 51, no. 9, pp. 1554–1560, Sep. 2003. [2] J. C. Ricklin and F. M. Davidson, “Atmospheric turbulence effects on a partially coherent Gaussian beam: Implications for free-space laser communication,” J. Opt. Soc. Amer., vol. 19, no. 9, pp. 1794–1802, 2002. [3] T. Ohtsuki, “Performance analysis of atmospheric optical PPM CDMA systems,” J. Lightw. Technol., vol. 21, no. 2, pp. 406–411, Feb. 2003. [4] D. Bushuev, D. Kedar, and S. Arnon, “Analyzing performance of nanosatellite cluster detector array receiver laser communication,” J. Lightw. Technol., vol. 21, no. 2, pp. 447–455, Feb. 2003. [5] J. Bannister, M. Gerla, and M. Kovaucevic, “An all-optical multifiber tree network,” J. Lightw. Technol., vol. 11, no. 5–6, pp. 997–1008, May–Jun. 1993. [6] S. Arnon, “The effects of atmospheric turbulence and building sway on optical wireless communication systems,” Opt. Lett., vol. 28, no. 2, pp. 129–131, 2003. [7] D. Kedar and S. Arnon, “Optical wireless communication through fog in the presence of pointing errors,” Appl. Opt., vol. 42, no. 24, pp. 1987–1993, 2003. [8] K. Ho and J. M. Kahn, “Methods for crosstalk measurement and reduction in dense WDM systems,” J. Lightw. Technol., vol. 14, no. 6, pp. 1127–1135, Jun. 1996. [9] M. Bhattacharya and T. Chattopadhyay, “Influence of adjacent channel interference on the frequency-modulated WDM optical communication system,” J. Lightw. Technol., vol. 17, no. 12, pp. 2516–2519, Dec. 1999. [10] E. Iannone, R. Sabella, M. Avattaneo, and G. D. Paolis, “Modeling of in-band crosstalk in WDM optical networks,” J. Lightw. Technol., vol. 17, no. 7, pp. 1135–1141, Jul. 1999. [11] Y. Shen, K. Lu, and W. Gu, “Coherent and incoherent crosstalk in WDM optical networks,” J. Lightw. Technol., vol. 17, no. 5, pp. 759–764, May 1999. [12] G. R. Hill et al., “A transport network layer based on optical network elements,” J. Lightw. Technol., vol. 11, no. 5–6, pp. 667–679, May–Jun. 1993. [13] P. Saengudomlert and M. Medard, “Guaranteeing the BER in transparent optical networks using OOK signaling,” IEEE J. Sel. Areas Commun., vol. 20, no. 4, pp. 786–799, May 2002. [14] I. T. Monroy and E. Tangdiongga, “Performance evaluation of optical cross-connects by saddlepoint approximation,” J. Lightw. Technol., vol. 16, no. 3, pp. 317–323, Mar. 1998. [15] R. M. Gagliardi, “Pulse-coded multiple access in space optical communications,” IEEE J. Sel. Areas Commun., vol. 13, no. 3, pp. 603–608, Apr. 1995. [16] B. R. Strickland, M. J. Lavan, E. Woodbridge, and V. Chan, “Effects of fog on the bit-error-rate of a free-space laser communication system,” Appl. Opt., vol. 38, no. 3, pp. 424–431, 1999. [17] E. J. McCartney, Optics of the Atmosphere. New York: Wiley, 1976. [18] R. A. McClatchey, R. W. Fenn, J. E. A. Selby, F. E. Volz, and J. S. Garing, “Optical properties of the atmosphere (revised),” Environ. Res. Papers, no. 354, p. 28, May 1971. [19] J. L. Buften and R. S. Iyer, “Continuous wave lidar measurement of atmospheric visibility,” Appl. Op., vol. 17, no. 2, pp. 265–271, 1978. [20] R. T. H. Collis and P. B. Russell, “Lidar measurement of particles and gases by elastic backscattering and differential absorption,” in Topics in Applied Physics, E. D. Hinkley, Ed. Berlin, Germany: SpringerVerlag, 1976, vol. 14, Laser monitoring of the atmosphere, ch. 4. [21] G. P. Agrawal, Fiber-Optic Communication Systems. New York: Wiley, 1997. [22] D. Kedar and S. Arnon, “Adaptive field-of-view receiver design for optical wireless communication through fog,” in Free Space Laser Communication Laser Imaging II Conf., Seattle, WA, Jul. 9, 2002.

2030

Debbie Kedar (S’02) received the M.A. degree in engineering science from Cambridge University, Cambridge, U.K., in 1982 and the M.Sc. degree from The Technion—Israel Institute of Technology, Haifa, Israel, in biomedical engineering in 1984. She is currently working toward the Ph.D. degree in the field of optical wireless communication at Ben-Gurion University of the Negev (BGU), Beer-Sheva, Israel. Current research interests include OWC in multiscattering channels, sensor networks.

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 23, NO. 6, JUNE 2005

Shlomi Arnon (SM’00) received the Ph.D. degree from Ben-Gurion University (BGU), Beer-Sheva, Israel. He was a Postdoctoral Associate (Fulbright Fellow) at the Laboratory for Information and Decision Systems (LIDS), Massachusetts Institute of Technology, Cambridge. He is currently a Faculty Member with the Electrical and Computer Engineering Department, BGU, and is the Founder of the Satellite and Wireless Communication Laboratory, which works intensively in the areas of laser satellite communication and terrestrial optical wireless communication systems. He was invited and sponsored by the U.S. Air Force to consult with scientists in laboratories in Rome and Albuquerque, NM. He consults regularly with start-up and well-established companies in the area of optical wireless communication and satellite communication. He delivered a workshop on laser satellite communication at NASA’s Jet Propulsion Laboratory, Pasadena, CA. Dr. Arnon is the IEEE Lasers & Electro-Optics Society (LEOS) Israeli Chapter Chair. He co-instructed a tutorial on laser satellite communication at the IEEE International Communication Conference (ICC) 2000 in New Orleans, LA. He was an Associate Editor for the Special Issue on Optical Wireless Communication of the Optical Society of America (OSA) Journal of Optical Networks. He recently delivered invited talks at The International Society for Optical Engineers (SPIE) International Symposium at Denver in 2004 and at the SPIE European Symposium on Optics and Photonics for Defence and Security in London, U.K., last year. He has also been invited to deliver a presentation at the 50th Annual SPIE meeting and International Symposium this summer in San Diego, CA.